Evaluator Observatory
The organizations that test frontier models before release, tracked as records: which lab's document names which evaluator, who pays whom and how much, who sits where and who moved, the laws and tenders building a market for them, and when each filing was last read. Every number links to its record and every record to the document it came from.
The labs say a model is safe to ship, and the sentence that carries the claim names an outside tester. This page keeps the ledger behind that sentence. It is built from a typed store of records, each one read off a document one of the parties published or a government filed, and nothing is on it that a record does not carry: a lab’s system card naming the evaluator, a funder’s grant table, a procurement notice, a team page, a bill-status page, an IRS return. The narrative that joins these records is on The Evaluator Alliances; the roster is on The AI Evaluators; the people are plotted in the revolving-door graph. This is the instrument panel: counts, dates, flows, and the receipt under each.
Three watchers feed a review queue behind it, none of them writing to the page: the IRS returns of each evaluator with an EIN, the developers’ own system-card pages for an evaluator newly named, and the bill-status pages of the California and Illinois statutes. A finding becomes a record only after a person has read the document.
▣ THE STORE
673 / 673 records in 8 tables
PAGE 2026-10-18
- 43/43evaluatorsas of 2026-10-23
- 156/156engagementsas of 2026-10-17
- 223/223moneyas of 2026-10-17
- 179/179seatsas of 2026-10-17
- 32/32movesas of 2026-10-17
- 9/9instrumentsas of 2026-10-17
- 13/13filingsas of 2026-10-17
- 18/18outcomesas of 2026-10-17
A tile's date is the newest read date of any record in that table. Every count on this page is a link to the rows behind it; every row carries the document it was read from. A cell the record cannot fill says so: not stated is a fact about the record, and each table ends with how many of its rows state each field.
▣ SLICE
673 / 673 records
Filters combine: any of the values ticked within a dimension, all of the dimensions at once. A table that has no column for a dimension you have set says so instead of showing rows that cannot match. The address bar carries the slice.
SLICE
edit ↑
Who tested whose model as of 2026-10-17
One engagement is one lab document naming one evaluator for one model: a system card that says the evaluator tested it, red-teamed it, built a test it ran, or was commissioned for a study. The cell counts them, across every developer with a document on record, and recounts under the slice above. Click a cell for the documents.
| Evaluator | OpenAI | Anthropic | Google DeepMind | Meta | xAI | Microsoft | Amazon | Mistral AI | Docs |
|---|---|---|---|---|---|---|---|---|---|
| UK AISI4 labs · dossier | · | · | · | · | 15 | ||||
| Gray Swan4 labs · dossier | · | · | · | · | 12 | ||||
| Irregular4 labs · dossier | · | · | · | · | 11 | ||||
| US CAISI4 labs · dossier | · | · | · | · | 11 | ||||
| Apollo Research4 labs · dossier | · | · | · | · | 7 | ||||
| Faculty4 labs · dossier | · | · | · | · | 6 | ||||
| SecureBio3 labs · dossier | · | · | · | · | · | 15 | |||
| Deloitte3 labs · dossier | · | · | · | · | · | 10 | |||
| Trajectory Labs3 labs · dossier | · | · | · | · | · | 7 | |||
| LatchBio2 labs · dossier | · | · | · | · | · | · | 9 | ||
| METR2 labs · dossier | · | · | · | · | · | · | 7 | ||
| Frontier Design2 labs · dossier | · | · | · | · | · | · | 4 | ||
| Gryphon Scientific2 labs · dossier | · | · | · | · | · | · | 4 | ||
| Scale AI2 labs · dossier | · | · | · | · | · | · | 3 | ||
| Vals AI2 labs · dossier | · | · | · | · | · | · | 3 | ||
| Epoch AI2 labs · dossier | · | · | · | · | · | · | 2 | ||
| Redwood Research1 labs · dossier | · | · | · | · | · | · | · | 8 | |
| Signature Science1 labs · dossier | · | · | · | · | · | · | · | 7 | |
| Andon Labs1 labs · dossier | · | · | · | · | · | · | · | 5 | |
| 10a Labs1 labs · dossier | · | · | · | · | · | · | · | 4 | |
| Charlemagne Labs1 labs · dossier | · | · | · | · | · | · | · | 2 | |
| MERIDIAN LABS1 labs · dossier | · | · | · | · | · | · | · | 2 | |
| Artificial Analysis1 labs · dossier | · | · | · | · | · | · | · | 1 | |
| Handshake AI1 labs · dossier | · | · | · | · | · | · | · | 1 | |
| Documents naming any evaluator | 21 | 102 | 7 | 18 | 3 | 2 | 1 | 2 | 156 |
Dated 155 of 156Model named 147 of 156Terms stated 75 of 156Finding quoted 62 of 156Developers with a document on record 8
Every engagement 156 rows, the document under each cell
Who pays whom as of 2026-10-17
One row is one amount as the record states it: a line in a funder's grant table, a contract notice, a Form 990 or 990-PF schedule, or the recipient's own disclosure. The ranked bars sum only rows that lead with a figure and name one funder and one recipient, in US dollars; sterling and euro contracts are totalled separately. A figure quoted inside prose (a valuation, a pledge, a fund's size, a round the funder joined) stays in the row's text, is marked as such, and is never summed. A record that names a funder without a figure says no amount published. A bar or a chip adds its funder or recipient to the slice.
Largest flows on record
Funders, by what the records sum to
Recipients
| Funder | Recipient | Amount, as recorded | When | Record | Source | Id |
|---|---|---|---|---|---|---|
| Coefficient Givingslice | Redwood Research slice | $9,420,000$9,420,000 (2022) | 2022 | grant database | Coefficient Giving; Coefficient Giving | 6db330fd5e |
| Coefficient Givingslice | Redwood Research slice | $10,700,000$10,700,000 (2022) | 2022 | grant database | Coefficient Giving; Coefficient Giving | e6188720ab |
| Coefficient Givingslice | Redwood Research slice | $5,300,000$5,300,000 (2023) | 2023 | grant database | Coefficient Giving; Coefficient Giving | 42609fe00b |
| Coefficient Givingslice | Redwood Research slice | $1,100,000$1,100,000 (2025) | 2025 | grant database | Coefficient Giving; Coefficient Giving | 59699f293d |
| Coefficient Givingslice | Redwood Research slice | $36,566,000$36,566,000 (Nov 2025) | Nov 2025 | grant database | Coefficient Giving; Coefficient Giving | 083f2a55d3 |
| Coefficient Givingslice | Redwood Research slice | $70,000,000stated (approximate)"more than $70 million over the next two years" (Sep 2026)over the next two years" | Sep 2026 | grant database | Coefficient Giving; Coefficient Giving | 883d059ed6 |
| Coefficient Givingslice | Epoch AI slice | $25,113,611Eight grants totalling $25,113,611 (Jun 2022 to Jul 2026), including $8,500,000 (Jun 2025)including $8,500,000 | Jun 2022 | party's own page | Epoch AI; Coefficient Giving | f97c5b81cf |
| Coefficient Givingslice | Epoch AI slice | $68,796,286$68,796,286 (Aug 2026), "a $68 million renewal over 3 years""a $68 million renewal over 3 years" | Aug 2026 | party's own page | Epoch AI; Coefficient Giving | 2f08995fb2 |
| Coefficient Givingslice | FAR.AI slice | $28,675,000$28,675,000 (Apr 2025) | Apr 2025 | grant database | Coefficient Giving | 3f45a5e003 |
| Coefficient Givingslice | FAR.AI slice | $6,650,000$6,650,000 (Apr 2025) | Apr 2025 | grant database | Coefficient Giving | a797d9271c |
| Coefficient Givingslice | FAR.AI slice | $2,420,253$2,420,253 (Oct to Nov 2025), after grants in Apr 2024 and Apr 2025after grants in | Nov 2025 | grant database | Coefficient Giving | c1a6fd3e88 |
| Coefficient Givingslice | Apollo Research slice | $1,535,480$1,535,480 (Jun 2023) | Jun 2023 | grant database | Coefficient Giving | 75364f8757 |
| Coefficient Givingslice | Apollo Research slice | $2,178,700$2,178,700 (May 2024) | May 2024 | grant database | Coefficient Giving | 1005449619 |
| Coefficient Givingslice | Apollo Research slice | $696,000$696,000 (May 2026) | May 2026 | grant database | Coefficient Giving | 2693517494 |
| Coefficient Givingslice | ARC slice | $265,000$265,000 (Mar 2022) | Mar 2022 | grant database | Coefficient Giving | 24020f82d0 |
| Coefficient Givingslice | ARC slice | $1,250,000$1,250,000 over two years (Nov 2022)over two years | Nov 2022 | grant database | Coefficient Giving | a151c338fc |
| Coefficient Givingslice | SecureBio slice | $1,420,937$1,420,937 (Mar 2023) | Mar 2023 | grant database | Coefficient Giving | 2d05c742c3 |
| Coefficient Givingslice | SecureBio slice | $570,000$570,000 (Jan 2024) | Jan 2024 | grant database | Coefficient Giving | f81fcc7f31 |
| Coefficient Givingslice | SecureBio slice | $3,430,000$3,430,000 (Mar 2025) | Mar 2025 | grant database | Coefficient Giving | 27af896435 |
| Coefficient Givingslice | SecureBio slice | $4,000,000$4,000,000 (Apr 2025) | Apr 2025 | grant database | Coefficient Giving | 9d61710bc5 |
| Coefficient Givingslice | SecureBio slice | $18,550$18,550 (Jun 2025) | Jun 2025 | grant database | Coefficient Giving | 4c3e0bd906 |
| Coefficient Givingslice | SecureBio slice | $37,000$37,000 (Jun 2025) | Jun 2025 | grant database | Coefficient Giving | 6475e9a6d1 |
| Coefficient Givingslice | Palisade Research slice | $1,680,000$1,680,000 (2024) | 2024 | grant database | Coefficient Giving | 13c6cb2752 |
| Coefficient Givingslice | Palisade Research slice | $2,123,463$2,123,463 (2025) | 2025 | grant database | Coefficient Giving | 063db329a9 |
| Coefficient Givingslice | RESOLUTION; Geoffrey irving | $160,000,000$160 million , "our largest ever AI safety grant""our largest ever AI safety grant" | not dated | grant database | Coefficient Giving; Resolution | 6e1b18b71f |
| Coefficient Givingslice | RESOLUTION; Geoffrey irving | no amount published"our sole funder to start" (2026) | 2026 | grant database | Coefficient Giving; Resolution | 3d98d14cdd |
| Coefficient Givingslice | GENERALITY LABS slice | $4,400,000" $4.4M in seed funding from Coefficient Giving" (2026) | 2026 | party's own page | Generality Labs | 91897decbe |
| Coefficient Givingslice | Stanford CRFM slice | figure in prose, not a stated paymentThe agent-benchmarks RFP "closed after funding $25 million in grants"The agent-benchmarks RFP "closed after funding | not dated | grant database | Coefficient Giving; 80,000 Hours | 683428a98a |
| Coefficient Givingslice | Stanford CRFM slice | no amount publishedCotra: "we funded a bunch of cool benchmarks, including Cybench" | not dated | grant database | Coefficient Giving; 80,000 Hours | 8eb72aa58c |
| Coefficient Givingslice | CENTER for AI SAFETY | $5,160,000$5,160,000 (Nov 2022) | Nov 2022 | grant database | Coefficient Giving | 44ee50efb9 |
| Coefficient Givingslice | CENTER for AI SAFETY | $1,866,559$1,866,559 exit grant (Oct 2023)exit grant | Oct 2023 | grant database | Coefficient Giving | 1034975d5f |
| Coefficient Givingslice | Georgetown CSET | $55,000,000$55,000,000 (Jan 2019) | Jan 2019 | grant database | Coefficient Giving | 01d8c8a3be |
| Coefficient Givingslice | Georgetown CSET | $38,920,000$38,920,000 (Aug 2021) | Aug 2021 | grant database | Coefficient Giving | 42ee1025ae |
| Coefficient Givingslice | RAND CORPORATION | $10,000,000$10,000,000, AI evaluation and testing (Sep 2025)AI evaluation and testing | Sep 2025 | grant database | Coefficient Giving | 9b52dc6dbb |
| Coefficient Givingslice | RAND CORPORATION | $6,000,000$6,000,000, AI security research (Nov 2025)AI security research | Nov 2025 | grant database | Coefficient Giving | efcfe3aa87 |
| Coefficient Givingslice | OPENAI | $30,000,000$30,000,000 general support (Mar 2017)general support | Mar 2017 | grant database | Coefficient Giving | a1d004c31e |
| Coefficient Givingslice | METR slice; Transluce slice; MERIDIAN LABS slice | no grant recordedNo entry under their own names in the funder's grant listing | not dated | grant database | Coefficient Giving | e0ba9e667c |
| Coefficient Givingslice | METR slice; Transluce slice; MERIDIAN LABS slice | no amount publishedMETR's predecessor ARC appears with the 2022 grants above | 2022 | grant database | Coefficient Giving | bf841e0386 |
| Survival and Flourishing Fund (Tallinn)slice | METR slice | $204,000$204,000 (2024) | 2024 | grant database | SFF-2024; SFF-2025; SFF-2026 | de27687f83 |
| Survival and Flourishing Fund (Tallinn)slice | METR slice | $548,000$548,000 total-funding line, "General support" (SFF-2025)total-funding line, "General support" | 2025 | grant database | SFF-2024; SFF-2025; SFF-2026 | c7e5f9b302 |
| Survival and Flourishing Fund (Tallinn)slice | METR slice | $2,201,000$2,201,000 (2026, $1,426,000 of it a matching pledge; funder column Jaan Tallinn) | 2026 | grant database | SFF-2024; SFF-2025; SFF-2026 | 9969c6a8a9 |
| Survival and Flourishing Fund (Tallinn)slice | ARC slice | $2,179,000$2,179,000 (2022-H2) | 2022-H2 | grant database | SFF-2022-H2; SFF-2023-H1; SFF-2024 | d77fe32fcc |
| Survival and Flourishing Fund (Tallinn)slice | ARC slice | $3,247,000$3,247,000 for the Evals Team (2023-H1)for the Evals Team | 2023-H1 | grant database | SFF-2022-H2; SFF-2023-H1; SFF-2024 | 88aee8f189 |
| Survival and Flourishing Fund (Tallinn)slice | ARC slice | $197,000$197,000 (2024) | 2024 | grant database | SFF-2022-H2; SFF-2023-H1; SFF-2024 | a40cdb0887 |
| Survival and Flourishing Fund (Tallinn)slice | Apollo Research slice | $490,000$490,000 (2023-H2) | 2023-H2 | grant database | SFF-2023-H2; SFF-2024; SFF-2026 | 8fdb77158b |
| Survival and Flourishing Fund (Tallinn)slice | Apollo Research slice | $251,000$251,000 (2024), via its then fiscal sponsor Rethink Prioritiesvia its then fiscal sponsor Rethink Priorities | 2024 | grant database | SFF-2023-H2; SFF-2024; SFF-2026 | bfdc0ce791 |
| Survival and Flourishing Fund (Tallinn)slice | Apollo Research slice | $315,000$315,000 (2026, Moskovitz $306,000 of it) | 2026 | grant database | SFF-2023-H2; SFF-2024; SFF-2026 | 1ec06b7e56 |
| Survival and Flourishing Fund (Tallinn)slice | SecureBio slice | $754,000$754,000 (SFF-2025) | 2025 | grant database | SFF-2025; SFF-2026 | 1163fe223c |
| Survival and Flourishing Fund (Tallinn)slice | SecureBio slice | $927,000$927,000 (SFF-2026) | 2026 | grant database | SFF-2025; SFF-2026 | fcf294adfd |
| Survival and Flourishing Fund (Tallinn; Moskovitz from 2026)slice | FAR.AI slice | $383,000$383,000 (2024) | 2024 | grant database | SFF-2024; SFF-2025; SFF-2026 | 02c1925c25 |
| Survival and Flourishing Fund (Tallinn; Moskovitz from 2026)slice | FAR.AI slice | $919,000$919,000 (2025) | 2025 | grant database | SFF-2024; SFF-2025; SFF-2026 | 53139cda55 |
| Survival and Flourishing Fund (Tallinn; Moskovitz from 2026)slice | FAR.AI slice | $875,000$875,000 (2026) | 2026 | grant database | SFF-2024; SFF-2025; SFF-2026 | 380cfe3557 |
| Survival and Flourishing Fund (Tallinn; Moskovitz from 2026)slice | FAR.AI slice | $580,000$580,000 (2024) to FAR AI, Inc. for the Safe AI Forumto FAR AI, Inc. for the Safe AI Forum | 2024 | grant database | SFF-2024; SFF-2025; SFF-2026 | e7dd072d42 |
| Survival and Flourishing Fund (Tallinn; Moskovitz from 2026)slice | Palisade Research slice | $586,000$586,000 (2024) | 2024 | grant database | SFF-2024; SFF-2025; SFF-2026; Palisade | e49380d76f |
| Survival and Flourishing Fund (Tallinn; Moskovitz from 2026)slice | Palisade Research slice | $1,133,000$1,133,000 (2025) | 2025 | grant database | SFF-2024; SFF-2025; SFF-2026; Palisade | e750a3e443 |
| Survival and Flourishing Fund (Tallinn; Moskovitz from 2026)slice | Palisade Research slice | $1,842,000$1,842,000 (2026) | 2026 | grant database | SFF-2024; SFF-2025; SFF-2026; Palisade | 82885c772f |
| Survival and Flourishing Fund (Tallinn; Moskovitz from 2026)slice | Palisade Research slice | figure in prose, not a stated paymenta 1:1 match up to $1.1 million in the Dec 2025 fundraisera 1:1 match up to | Dec 2025 | grant database | SFF-2024; SFF-2025; SFF-2026; Palisade | 24c54ae0e8 |
| Survival and Flourishing Fund; Jaan TallinnJaan tallinnslice · slice | Epoch AI slice | $195,000$195,000 (May 2026) | May 2026 | party's own page | Epoch AI; SFF-2026 | bb9e1db106 |
| Jaan Tallinn (personally)slice | Epoch AI slice | $600,000$600,000 (Jan 2025) | Jan 2025 | party's own page | Epoch AI; SFF-2026 | e6b88d3b90 |
| Jaan Tallinn (personally)slice | Epoch AI slice | $646,000SFF-2026 recommends $646,000 in all (Moskovitz $542,000 of it)SFF- recommends | 2026 | party's own page | Epoch AI; SFF-2026 | aafd049fe8 |
| Jaan Tallinnslice | FRONTIER MODEL FORUM | no amount publishedNamed supporter, no amount | not dated | party's own page | Frontier Model Forum | 2a25db90a4 |
| Schmidt Sciencesslice | METR slice | no amount publishedNamed funder, no amount | not dated | party's own page | METR; METR | 0f7d6ee257 |
| Schmidt Sciencesslice | METR slice | figure in prose, not a stated paymentpart of "around $71 million " of commitments in six months (Aug 2026)" of commitments in six months | Aug 2026 | party's own page | METR; METR | 2b32914e83 |
| Schmidt Sciencesslice | Epoch AI slice | no amount publishedNamed donor (Dec 2025), no amount | Dec 2025 | party's own page | Epoch AI | 08a0bf8e3c |
| Schmidt Sciencesslice | UK AISI slice | figure in prose, not a stated paymentNamed funder of the "over £27 million " fund, with OpenAI, Anthropic, Microsoft and AWS" fund, with OpenAI, Anthropic, Microsoft and AWS | not dated | party's own page | Alignment Project | 2515f554c5 |
| Schmidt Sciencesslice | FRONTIER MODEL FORUM | no amount publishedNamed supporter, no amount | not dated | party's own page | Frontier Model Forum | b3556a6482 |
| Schmidt Sciencesslice | Adam gleave; FAR.AI slice; Percy liang; Stanford CRFM slice | no amount publishedAI2050 Early Career Fellow (Nov 2025) | Nov 2025 | party's own page | FAR.AI; CRFM | 039a88b316 |
| Schmidt Sciencesslice | Adam gleave; FAR.AI slice; Percy liang; Stanford CRFM slice | no amount publishedAI2050 Senior Fellow (2022) | 2022 | party's own page | FAR.AI; CRFM | 4a18839af9 |
| Schmidt Sciencesslice | Adam gleave; FAR.AI slice; Percy liang; Stanford CRFM slice | no amount publishedSchmidt Futures a named CRFM supporter | not dated | party's own page | FAR.AI; CRFM | 86a658e98e |
| Eric Schmidt (personally)slice | ROOST (ROBUST OPEN ONLINE SAFETY TOOLS) | no amount publishedFounding partner, with Discord, OpenAI, Google, Roblox and four foundations | not dated | party's own page | PR Newswire | 7e31c90f97 |
| Eric Schmidt (personally)slice | ROOST (ROBUST OPEN ONLINE SAFETY TOOLS) | figure in prose, not a stated payment"more than $27 million " raised in all, no per-funder amount" raised in all, no per-funder amount | not dated | party's own page | PR Newswire | ad0613db56 |
| Patrick J. McGovern Foundationslice | ROOST (ROBUST OPEN ONLINE SAFETY TOOLS); Stanford CRFM slice | no amount publishedNamed supporter of the Fund | not dated | party's own page | Frontier Model Forum; PR Newswire; CRFM | 29192c76e7 |
| Patrick J. McGovern Foundationslice | ROOST (ROBUST OPEN ONLINE SAFETY TOOLS); Stanford CRFM slice | no amount publishedROOST founding partner | not dated | party's own page | Frontier Model Forum; PR Newswire; CRFM | 244aa0bd09 |
| Patrick J. McGovern Foundationslice | ROOST (ROBUST OPEN ONLINE SAFETY TOOLS); Stanford CRFM slice | no amount publishedCRFM supporter | not dated | party's own page | Frontier Model Forum; PR Newswire; CRFM | 3c911903b4 |
| Patrick J. McGovern Foundationslice | ROOST (ROBUST OPEN ONLINE SAFETY TOOLS); Stanford CRFM slice | no amount publishedno amounts | not dated | party's own page | Frontier Model Forum; PR Newswire; CRFM | 9c1ae442eb |
| Frontier Model Forum AI Safety Fund (Anthropic, Google, Microsoft, OpenAI, with Packard, Schmidt Sciences, Tallinn, McGovern)slice | Apollo Research slice; FAR.AI slice; SecureBio slice | no amount publishedDecember 2025 cohort | December 2025 | party's own page | Frontier Model Forum | a0805da5d0 |
| Frontier Model Forum AI Safety Fund (Anthropic, Google, Microsoft, OpenAI, with Packard, Schmidt Sciences, Tallinn, McGovern)slice | Apollo Research slice; FAR.AI slice; SecureBio slice | no amount publishedamounts not stated in the announcement | not dated | party's own page | Frontier Model Forum | b44d635f47 |
| OpenAI Foundationslice | SecureBio slice | $17,200,000$17.2M to Detection, restricted (Aug 2026)to Detection, restricted | Aug 2026 | party's own page | SecureBio | c2832b57e2 |
| OpenAIslice | RESOLUTION | $10,000,000in kind, stated$10 million in API credits (Aug 2026) | Aug 2026 | party's own page | Resolution | af1d05e00e |
| OpenAIslice | Epoch AI slice | no amount publishedCommissioned and funded FrontierMath's 300-problem core | not dated | party's own page | Epoch AI; Sevilla | 2555c5299b |
| OpenAIslice | Epoch AI slice | no amount published"OpenAI purchased access to the verifiers" (Apr 2026) | Apr 2026 | party's own page | Epoch AI; Sevilla | e85d2d0a4a |
| OpenAIslice | METR slice; Redwood Research slice | $400,000in kind, stated (approximate)Free API credits of roughly $400K for the Hugging Face investigationfor the Hugging Face investigation | not dated | party's own page | Redwood Research | 15ba3688d2 |
| OpenAIslice | METR slice; Redwood Research slice | no amount publishedno payment | not dated | party's own page | Redwood Research | fac806a2d5 |
| OpenAI employees' personal holdingsslice | Transluce slice | share of revenue, no amount6% of fiscal 2025 revenue | 2025 | party's own page | Transluce | 22ab6796c0 |
| Anthropic employees' personal holdingsslice | Transluce slice | share of revenue, no amount32% of fiscal 2025 revenue | 2025 | party's own page | Transluce | 44f5ad8239 |
| OpenAI, Anthropic, Microsoft, AWS, Schmidt SciencesANTHROPIC · MICROSOFT · SCHMIDT SCIENCESslice · slice · slice · slice · slice | UK AISI slice | £27,000,000stated (approximate)"over £27 million ""over | not dated | party's own page | Alignment Project | 2be668d22f |
| OpenAI, Anthropic, Google DeepMind, Meta, AmazonANTHROPIC · GOOGLE DEEPMIND · METAslice · slice · slice · slice · slice | METR slice | no amount published"access and tokens," no cash | not dated | party's own page | METR; METR | 9fe2dac358 |
| OpenAI, Anthropic, Google DeepMind, Meta, AmazonANTHROPIC · GOOGLE DEEPMIND · METAslice · slice · slice · slice · slice | METR slice | no amount publishedthe Opus 5.5 evaluation "under an unpaid agreement" | not dated | party's own page | METR; METR | 3517e823a8 |
| The Audacious Projectslice | METR slice | $17,000,000stated (approximate)"Approximately $17 million " of a $38 million Canary award shared with RAND (Oct 2024)" of a $38 million Canary award shared with RAND | Oct 2024 | party's own page | METR | 9c5401b3e9 |
| UK Department for Science, Innovation and Technology (and DESNZ)slice | Trail of Bits slice | £255,600£255,600 | not dated | contract notice | Find a Tender; Find a Tender; Contracts Finder; Contracts Finder; Contracts Finder; Find a Tender, variation | 7f7837322d |
| UK Department for Science, Innovation and Technology (and DESNZ)slice | Gryphon Scientific slice | £267,618£267,618 (varied to £392,486.78, March 2024) | March 2024 | contract notice | Find a Tender; Find a Tender; Contracts Finder; Contracts Finder; Contracts Finder; Find a Tender, variation | bb860f8d21 |
| UK Department for Science, Innovation and Technology (and DESNZ)slice | Irregular slice | £195,434£195,434 | not dated | contract notice | Find a Tender; Find a Tender; Contracts Finder; Contracts Finder; Contracts Finder; Find a Tender, variation | e376addb9f |
| UK Department for Science, Innovation and Technology (and DESNZ)slice | Gray Swan slice | £128,770£128,770 (Jan 2024) | Jan 2024 | contract notice | Find a Tender; Find a Tender; Contracts Finder; Contracts Finder; Contracts Finder; Find a Tender, variation | e18da52332 |
| UK Department for Science, Innovation and Technology (and DESNZ)slice | Gray Swan slice | £440,000£440,000 to Gray Swan (Aug 2024)to Gray Swan | Aug 2024 | contract notice | Find a Tender; Find a Tender; Contracts Finder; Contracts Finder; Contracts Finder; Find a Tender, variation | d1d7123d3d |
| UK AI Security Instituteslice | METR slice | no amount publishedMETR lists the Institute as a donor | not dated | party's own page | METR; Epoch AI; Manifund; Generality Labs | 2a3ab21c74 |
| UK AI Security Instituteslice | Epoch AI slice | no amount publishedEpoch lists UK DSIT as a paying partner | not dated | party's own page | METR; Epoch AI; Manifund; Generality Labs | f46738332c |
| UK AI Security Instituteslice | GENERALITY LABS slice | $1,000,000stated (approximate)the Inspect Evals project "has been contracted with the UK AISI since January '25" and counted more than $1M in contracts with the Institute, Epoch and METR before its spin-offthe Inspect Evals project "has been contracted with the UK AISI since January '25" and counted more than | not dated | party's own page | METR; Epoch AI; Manifund; Generality Labs | bf946cf3f4 |
| European AI Officeslice | METR slice | no amount publishedMETR: a technical-assistance contract, "a small part of our income" | not dated | party's own page | METR; Epoch AI; Transluce; FAR.AI | aa7cb6963c |
| European AI Officeslice | Epoch AI slice | no amount publishedEpoch: a paying partner (2026) | 2026 | party's own page | METR; Epoch AI; Transluce; FAR.AI | f59b15efd6 |
| European AI Officeslice | Transluce slice | no amount publishedTransluce: "Serving as a contractor to the EU AI Office" | not dated | party's own page | METR; Epoch AI; Transluce; FAR.AI | 644f0e05c1 |
| European AI Officeslice | FAR.AI slice | no amount publishedFAR.AI: leads the consortium selected under tender EC-CNECT/2025/OP/0032, Lot 1 | 2025 | party's own page | METR; Epoch AI; Transluce; FAR.AI | a0e64c1b24 |
| METRslice | Epoch AI slice | no amount publishedA 2025 software-engineering benchmark engagement, funded by METR | 2025 | party's own page | Epoch AI | 7c3cc5bec8 |
| Nat Friedmanslice | Epoch AI slice; MECHANIZE slice; Artificial Analysis slice | no amount publishedFunded a pilot of FrontierMath | not dated | party's own page | Epoch AI; Mechanize; Artificial Analysis; AIUC, Series A; Fortune, Oct 8 2025; Meta memo via CNBC | 16a5ebe3a8 |
| Nat Friedmanslice | Epoch AI slice; MECHANIZE slice; Artificial Analysis slice | no amount publishedfirst-named backer of Mechanize | not dated | party's own page | Epoch AI; Mechanize; Artificial Analysis; AIUC, Series A; Fortune, Oct 8 2025; Meta memo via CNBC | cf9b872cbf |
| Nat Friedmanslice | Epoch AI slice; MECHANIZE slice; Artificial Analysis slice | no amount publishedfirst-named of the eight backers on Artificial Analysis's About page, no amount stated | not dated | party's own page | Epoch AI; Mechanize; Artificial Analysis; AIUC, Series A; Fortune, Oct 8 2025; Meta memo via CNBC | 5e85e6667c |
| Nat Friedmanslice | Epoch AI slice; MECHANIZE slice; Artificial Analysis slice | $55,000,000named on AIUC's team page among its $55 million investors, and AIUC's Series A post counts "our $15 million seed round from NFDG"investors, and AIUC's Series A post counts "our $15 million seed round from NFDG" | not dated | party's own page | Epoch AI; Mechanize; Artificial Analysis; AIUC, Series A; Fortune, Oct 8 2025; Meta memo via CNBC | 4aef243df2 |
| Nat Friedmanslice | Epoch AI slice; MECHANIZE slice; Artificial Analysis slice | no amount publishedFortune names him among the seeders of Situational Awareness LP | not dated | party's own page | Epoch AI; Mechanize; Artificial Analysis; AIUC, Series A; Fortune, Oct 8 2025; Meta memo via CNBC | 0291b6507f |
| Nat Friedmanslice | Epoch AI slice; MECHANIZE slice; Artificial Analysis slice | no amount publishedsince June 2025 co-lead of Meta Superintelligence Labs for AI products and applied research, per Meta's 30 June 2025 memo | June 2025 | party's own page | Epoch AI; Mechanize; Artificial Analysis; AIUC, Series A; Fortune, Oct 8 2025; Meta memo via CNBC | 835d93c2fa |
| In-Q-Telslice | Dreadnode slice | figure in prose, not a stated paymentParticipant in the $14 million Series A led by Decibel, 25 February 2025Series A led by Decibel | 25 February 2025 | party's own page | Dreadnode; IQT | dfa5abb0df |
| In-Q-Telslice | Dreadnode slice | no amount publishedIQT calls itself "an independent, not-for-profit strategic investment firm" | not dated | party's own page | Dreadnode; IQT | bbb140866d |
| In-Q-Telslice | Dreadnode slice | $114,227,920its return for the tax year to 31 March 2025 reports $114,227,920 of itsof its | 31 March 2025 | party's own page | Dreadnode; IQT | e88bfba7ea |
| In-Q-Telslice | Dreadnode slice | $132,130,603$132,130,603 revenue as government grantsrevenue as government grants | not dated | party's own page | Dreadnode; IQT | 81a84e354c |
| Andreessen Horowitz (a16z)slice | Vals AI slice | figure in prose, not a stated payment$40M Series A at a $400M valuation (August 2026)Series A at a $400M valuation | August 2026 | party's own page | Vals AI; Arena; Arena, seed; Braintrust; Promptfoo; OpenAI; FEC | 6542e7a43a |
| Andreessen Horowitz (a16z)slice | Vals AI slice; Arena slice | no amount publishedArena says it "raised funding from a16z and The University of California" | not dated | party's own page | Vals AI; Arena; Arena, seed; Braintrust; Promptfoo; OpenAI; FEC | 7c90a4560f |
| Andreessen Horowitz (a16z)slice | Arena slice | figure in prose, not a stated payment$100M seed (27 May 2025)seed | 27 May 2025 | party's own page | Vals AI; Arena; Arena, seed; Braintrust; Promptfoo; OpenAI; FEC | 4386776373 |
| Andreessen Horowitz (a16z)slice | Vals AI slice; Arena slice | figure in prose, not a stated paymentled Braintrust's $36M Series A (8 October 2024)Series A | 8 October 2024 | party's own page | Vals AI; Arena; Arena, seed; Braintrust; Promptfoo; OpenAI; FEC | 98eb887980 |
| Andreessen Horowitz (a16z)slice | Vals AI slice; Arena slice | no amount publisheda participant in Promptfoo's Series A (29 July 2025), since acquired by OpenAI | 29 July 2025 | party's own page | Vals AI; Arena; Arena, seed; Braintrust; Promptfoo; OpenAI; FEC | 6cc58353c7 |
| Andreessen Horowitz (a16z)slice | Vals AI slice; Arena slice | no amount publishedco-led OpenAI's March 2026 round alongside SoftBank | March 2026 | party's own page | Vals AI; Arena; Arena, seed; Braintrust; Promptfoo; OpenAI; FEC | dab4d02d57 |
| Andreessen Horowitz (a16z)slice | Vals AI slice; Arena slice | $100,000,000A16Z Capital Management, Marc Andreessen and Ben Horowitz gave $100,000,000 to the Leading the Future super PAC on two dates (FEC)to the Leading the Future super PAC on two dates (FEC | not dated | party's own page | Vals AI; Arena; Arena, seed; Braintrust; Promptfoo; OpenAI; FEC | 60729df417 |
| Andreessen Horowitz (a16z)slice | Vals AI slice; Arena slice | no amount publishedno stake size published | not dated | party's own page | Vals AI; Arena; Arena, seed; Braintrust; Promptfoo; OpenAI; FEC | 6c7b9d3282 |
| Fathom; Halcyon Futures; Coefficient Giving; Geoff Ralston (also a METR supporter); the AI Underwriting Company; Craig Falls; Good Forever Foundation; Sympatico Ventures; non-executive employees and alumni of frontier AI companiesHALCYON FUTURES · OPEN PHILANTHROPY (COEFFICIENT GIVING) · AIUCslice · slice · slice · slice · slice · slice · slice · slice · slice | AVERI slice | no amount publishedNo amounts published | not dated | party's own page | AVERI, About | 3ecd726491 |
| Fathom; Halcyon Futures; Coefficient Giving; Geoff Ralston (also a METR supporter); the AI Underwriting Company; Craig Falls; Good Forever Foundation; Sympatico Ventures; non-executive employees and alumni of frontier AI companiesHALCYON FUTURES · OPEN PHILANTHROPY (COEFFICIENT GIVING) · AIUCslice · slice · slice · slice · slice · slice · slice · slice · slice | AVERI slice | no amount publishedAPI credits from Amazon, Anthropic, Google DeepMind, Microsoft, OpenAI and Thinking Machines Lab | not dated | party's own page | AVERI, About | bd3a0dae0d |
| Coefficient Givingslice | MATS RESEARCH | $3,382,029$3,382,029 (Feb 2025) | Feb 2025 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | b82cf3527e |
| Coefficient Givingslice | MATS RESEARCH | $660,000$660,000 (May 2025) to MATS Researchto MATS Research | May 2025 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | 2d34baa7a1 |
| Coefficient Givingslice | MATS RESEARCH | $2,047,268$2,047,268 (Mar 2023) | Mar 2023 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | 1c95c2e210 |
| Coefficient Givingslice | MATS RESEARCH | $2,641,368$2,641,368 (Mar 2024) to BERIto BERI | Mar 2024 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | 167f4b540d |
| Coefficient Givingslice | MATS RESEARCH | $732,631$732,631 (Mar 2024), | Mar 2024 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | 9b963acf3d |
| Coefficient Givingslice | MATS RESEARCH | $2,321,550$2,321,550 (Jun 2025) | Jun 2025 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | 9ad4afb536 |
| Coefficient Givingslice | MATS RESEARCH | $709,567$709,567 (Jul 2025) to AI Safety Support for the program. MATS's own 2025 cash lines: Coefficient Givingto AI Safety Support for the program. MATS's own cash lines: Coefficient Giving | Jul 2025 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | 6f183f8747 |
| Coefficient Givingslice | MATS RESEARCH | $11,823,980$11,823,980 (Jul) (Sep 2025) | Sep 2025 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | 6812f7633a |
| Coefficient Givingslice | MATS RESEARCH | $3,629,280$3,629,280 (Sep) via two intermediaries (Sep 2025)via two intermediaries | Sep 2025 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | 7e5bc27e3c |
| Good Ventures Foundationslice | MATS RESEARCH | $1,520,150$1,520,150 | not dated | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | 7b53cc6121 |
| Good Ventures Foundationslice | MATS RESEARCH | $3,038,300$3,038,300 (May) | not dated | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | bf6b47d1ec |
| Good Ventures Foundationslice | MATS RESEARCH | figure in prose, not a stated paymenttogether $20,011,710 of $20,891,404. Good Ventures Foundation's Form 990-PF for the tax year to 30 Jun 2025 prints the same two amounts to MATS Research Inc for the Winter 24-25 cohortof $20,891,404. Good Ventures Foundation's Form 990-PF for the tax year to prints the same two amounts to MATS Research Inc for the Winter 24-25 cohort | 30 Jun 2025 | grant database | Coefficient Giving; MATS, Transparency; ProPublica, Good Ventures FY2024 Form 990-PF | c3a25bae47 |
| Survival and Flourishing Fund (Tallinn)slice | MATS RESEARCH | $639,000$639,000 (SFF-2024, MATS Research, Inc., 'General support') | 2024 | grant database | SFF-2024; SFF-2025; SFF-2026 | 3b0215a8ee |
| Survival and Flourishing Fund (Tallinn)slice | MATS RESEARCH | $289,000$289,000 (SFF-2025, Main, 'General support') | 2025 | grant database | SFF-2024; SFF-2025; SFF-2026 | c3ac0cf6df |
| Survival and Flourishing Fund (Tallinn)slice | MATS RESEARCH | no grant recordednot in the 2026 table | 2026 | grant database | SFF-2024; SFF-2025; SFF-2026 | 9cd0464984 |
| Longview Philanthropyslice | MATS RESEARCH | $1,052,000$1,052,000 on Longview Philanthropy USA Inc's TY2025 Schedule I "to fund an expansion of a training program for early-career AI safety researchers" | not dated | 990 | ProPublica, Longview Philanthropy USA TY2025 Schedule I; MATS, AI Safety Talent Needs in 2026; MATS, Transparency | ccc844f21c |
| Longview Philanthropyslice | MATS RESEARCH | no amount publishedthanked in MATS's March 2026 report among donors 'without whose donations we would be unable to run upcoming programs' | March 2026 | 990 | ProPublica, Longview Philanthropy USA TY2025 Schedule I; MATS, AI Safety Talent Needs in 2026; MATS, Transparency | b13ba0bd99 |
| Longview Philanthropyslice | MATS RESEARCH | no grant recordedno Longview line on MATS's transparency page | not dated | 990 | ProPublica, Longview Philanthropy USA TY2025 Schedule I; MATS, AI Safety Talent Needs in 2026; MATS, Transparency | 2c66200689 |
| Ryan Kidd (Manifund regrant budget)slice | Transluce | $10,000$10,000 (25 Mar 2026) | 25 Mar 2026 | grant database | Manifund, Ryan Kidd | 06573c187d |
| Ryan Kidd (Manifund regrant budget)slice | Apart Research | $5,000$5,000 (11 Jan 2024) | 11 Jan 2024 | grant database | Manifund, Ryan Kidd | 831f99ebb8 |
| Ryan Kidd (Manifund regrant budget)slice | Apart Research | $5,000$5,000 (18 Jun 2025) | 18 Jun 2025 | grant database | Manifund, Ryan Kidd | c5ad6015d4 |
| Ryan Kidd (Manifund regrant budget)slice | Institute for AI Policy and Strategy | $5,000$5,000 (9 Jan 2024) | 9 Jan 2024 | grant database | Manifund, Ryan Kidd | 745c46c519 |
| Ryan Kidd (Manifund regrant budget)slice | PIBBSS | $25,000$25,000 (13 Sep 2024) | 13 Sep 2024 | grant database | Manifund, Ryan Kidd | d65111e862 |
| Ryan Kidd (Manifund regrant budget)slice | BASE | $5,000$5,000 (3 Aug 2026) | 3 Aug 2026 | grant database | Manifund, Ryan Kidd | 72e7650390 |
| Ryan Kidd (Manifund regrant budget)slice | Inspect Evals | $19,997$19,997 (29 Oct 2025) | 29 Oct 2025 | grant database | Manifund, Ryan Kidd | 29cfe17e46 |
| Anthropicslice | MATS RESEARCH | $35,000$35,000 (late 2024) | 2024 | party's own page | MATS, Transparency | 3d55220a44 |
| Googleslice | MATS RESEARCH | $100,000$100,000 (Nov 2025) | Nov 2025 | party's own page | MATS, Transparency | 6a83962273 |
| Googleslice | MATS RESEARCH | no amount publishedThinking Machines research grants (Nov 2025), no value stated | Nov 2025 | party's own page | MATS, Transparency | a0aad606e8 |
| OpenAIslice | MATS RESEARCH | $120,000$120,000 (Mar 2026) | Mar 2026 | party's own page | MATS, Transparency | c443b70f1b |
| Microsoftslice | MICROSOFT | no amount published"unrestricted funding to 18 university labs across six continents" | not dated | party's own page | Microsoft RAI report | 77c7669bbf |
| Microsoftslice | FRONTIER MODEL FORUM | no amount published"wetlab" research grants led by the Forum, no amount stated | not dated | party's own page | Microsoft RAI report | 8eb45b4811 |
| Good Ventures Foundationslice | RAND CORPORATION | $2,500,000$2,500,000 (30 Jun 2025)general support | 30 Jun 2025 | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | a5e7071c5e |
| Good Ventures Foundationslice | RAND CORPORATION | $2,000,000$2,000,000 (30 Jun 2025)general support | 30 Jun 2025 | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | d8d1c044d8 |
| Good Ventures Foundationslice | RAND CORPORATION | $666,667$666,667, each general support (tax year to 30 Jun 2025)general support | 30 Jun 2025 | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | f9a1a3da09 |
| Good Ventures Foundationslice | GOVAI (CENTRE for the GOVERNANCE of AI) | $2,500,000$2,500,000 (30 Jun 2025)general support | 30 Jun 2025 | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | c60181e289 |
| Good Ventures Foundationslice | GOVAI (CENTRE for the GOVERNANCE of AI) | $495,207$495,207, each general support (tax year to 30 Jun 2025)general support | 30 Jun 2025 | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | 63a87f0f59 |
| Good Ventures Foundationslice | SecureBio slice | $1,400,000$1,400,000 general support (30 Jun 2025)general support | 30 Jun 2025 | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | 321dab7e04 |
| Good Ventures Foundationslice | SecureBio slice | $230,000$230,000 for a public health early warning system, paid (tax year to 30 Jun 2025)for a public health early warning system | 30 Jun 2025 | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | 680f9117b1 |
| Good Ventures Foundationslice | SecureBio slice | $2,600,000$2,600,000 general support approved for future paymentgeneral support | not dated | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | 38515a7641 |
| Good Ventures Foundationslice | FAR.AI slice | $676,180$676,180 (advance public knowledge of AI alignment and safety)advance public knowledge of AI alignment and safety | not dated | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | 169e92125d |
| Good Ventures Foundationslice | Epoch AI slice | $70,000$70,000 (improvements to the FrontierMath benchmark)improvements to the FrontierMath benchmark | not dated | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | afb66f19f8 |
| Good Ventures Foundationslice | Forecasting Research Institute | $358,850$358,850 (forecasting evaluations of advanced AI systems), tax year to 30 Jun 2025forecasting evaluations of advanced AI systems), tax year to | 30 Jun 2025 | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | 24c298e17b |
| Good Ventures Foundationslice | Irregular slice | $2,266,667$2,266,667, developing software tools, products and analysis focused on global security (tax year to 30 Jun 2025)developing software tools, products and analysis focused on global security | 30 Jun 2025 | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | 6bf68ee717 |
| Good Ventures Foundationslice | Irregular slice | no amount publishedthe return does not name Irregular | not dated | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF | 74571eeb90 |
| Good Ventures Foundation (amounts equal to Coefficient Giving records)slice | AI Safety Support Ltd (MATS winter 2024-25 cohort) | $1,004,850$1,004,850 | not dated | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF; Coefficient Giving; Coefficient Giving | 0981d871ff |
| Good Ventures Foundation (amounts equal to Coefficient Giving records)slice | AI Safety Support Ltd (MATS winter 2024-25 cohort) | $1,316,700$1,316,700 = $2,321,550, the pair of grants Coefficient Giving records for the same cohort= $2,321,550, the pair of grants Coefficient Giving records for the same cohort | not dated | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF; Coefficient Giving; Coefficient Giving | 30a94c30a2 |
| Good Ventures Foundation (amounts equal to Coefficient Giving records)slice | Princeton University (AI R&D capabilities benchmark) | $863,143$863,143, the amount of Coefficient Giving's Princeton AI R&D benchmark grantthe amount of Coefficient Giving's Princeton AI R&D benchmark grant | not dated | 990-PF | ProPublica, Good Ventures FY2024 Form 990-PF; Coefficient Giving; Coefficient Giving | 196b5e1296 |
| Google.org; Heising-Simons Foundationslice · slice | Humane Intelligence slice | no amount publishedBoth named funders on the nonprofit's own pages, no amounts: Google.org for the original bias-bounty program and the open-source red-teaming application ('fully funded', Dec 2025) | Dec 2025 | party's own page | Humane Intelligence, 2025 Impact Report; Humane Intelligence, Bias Bounties at Scale; ProPublica | 3e31fa497b |
| Heising-Simons Foundationslice | Humane Intelligence slice | figure in prose, not a stated payment$100,000, program-service revenueprogram-service revenue | not dated | party's own page | Humane Intelligence, 2025 Impact Report; Humane Intelligence, Bias Bounties at Scale; ProPublica | fb943c87da |
| Heising-Simons Foundationslice | Humane Intelligence slice | $372,175$372,175 | not dated | party's own page | Humane Intelligence, 2025 Impact Report; Humane Intelligence, Bias Bounties at Scale; ProPublica | 307f6ec1ef |
| Survival and Flourishing Fund (source Jaan Tallinn)slice | Collective Intelligence Project slice | $300,000$300,000 (SFF-2023-H1) | 2023-H1 | grant database | SFF, Recommendations | 2056d1fd64 |
| Survival and Flourishing Fund (source Jaan Tallinn)slice | Collective Intelligence Project slice | $500,000$500,000 (2024) | 2024 | grant database | SFF, Recommendations | a17a60b8d5 |
| Survival and Flourishing Fund (source Jaan Tallinn)slice | Collective Intelligence Project slice | $500,000$500,000 (Initiative Committee 2024, one through Radicalxchange Ltd. for CIP's Global Representation/Constitutions Project) | 2024 | grant database | SFF, Recommendations | 25e1fc0636 |
| Survival and Flourishing Fund (source Jaan Tallinn)slice | Collective Intelligence Project slice | no grant recordednot named in the 2024, 2025 or 2026 main-round tables | 2024 | grant database | SFF, Recommendations | 033ccec3b4 |
| Future of Life Institute; Google.org; Omidyar Network; Robert Wood Johnson Foundation; Ford Foundation; Fast Forwardslice · slice · slice · slice · slice · slice | Collective Intelligence Project slice | no amount publishedNo amounts published | not dated | party's own page | CIP, 2025 Annual Report; CIP, 2024 Annual Report; ProPublica | 5cb3b4384a |
| Future of Life Institute; Google.org; Omidyar Network; Robert Wood Johnson Foundation; Ford Foundation; Fast Forwardslice · slice · slice · slice · slice · slice | Collective Intelligence Project slice | figure in prose, not a stated paymentFY2024 contributions $1,907,953 and FY2025 contributionsFY2025 contributions | not dated | party's own page | CIP, 2025 Annual Report; CIP, 2024 Annual Report; ProPublica | 57034a2c15 |
| Future of Life Institute; Google.org; Omidyar Network; Robert Wood Johnson Foundation; Ford Foundation; Fast Forwardslice · slice · slice · slice · slice · slice | Collective Intelligence Project slice | $98,460$98,460 (Form 990 summary)Form 990 summary | not dated | party's own page | CIP, 2025 Annual Report; CIP, 2024 Annual Report; ProPublica | d5940b3bff |
| Future of Life Institute; Google.org; Omidyar Network; Robert Wood Johnson Foundation; Ford Foundation; Fast Forwardslice · slice · slice · slice · slice · slice | Collective Intelligence Project slice | no amount publishedAnthropic, OpenAI, Mistral AI, Microsoft Research and Fathom listed as organizational partners | not dated | party's own page | CIP, 2025 Annual Report; CIP, 2024 Annual Report; ProPublica | b92b78b5f5 |
| Coefficient Givingslice | Princeton Holistic Agent Leaderboard slice | $863,143$863,143 (Nov 2024) to Princeton for an AI R&D benchmark led by Arvind Narayanan, via the agent-benchmarks request for proposalsto Princeton for an AI R&D benchmark led by Arvind Narayanan, via the agent-benchmarks request for proposals | Nov 2024 | grant database | Coefficient Giving; HAL | b8151eea2e |
| Coefficient Givingslice | Princeton Holistic Agent Leaderboard slice | $1,045,620$1,045,620 (Jul 2024) to Princeton for Karthik Narasimhan's software-engineering benchmarkto Princeton for Karthik Narasimhan's software-engineering benchmark | Jul 2024 | grant database | Coefficient Giving; HAL | 356ee01ead |
| Coefficient Givingslice | Princeton Holistic Agent Leaderboard slice | no amount publishedneither listing names HAL, whose own page names Coefficient Giving as a funder | not dated | grant database | Coefficient Giving; HAL | c45920c4eb |
| Schmidt Sciencesslice | METR slice; Princeton Holistic Agent Leaderboard slice | no amount publishedHAL: named funder on its own page | not dated | party's own page | HAL; Schmidt Sciences | bb7cf8ceb4 |
| Schmidt Sciencesslice | METR slice; Princeton Holistic Agent Leaderboard slice | no amount publishedSchmidt's AI Safety Science program (27 projects, 11 Feb 2025) names Arvind Narayanan among its cohort, with "API access from OpenAI" | 11 Feb 2025 | party's own page | HAL; Schmidt Sciences | e9d4874644 |
| Schmidt Sciencesslice | METR slice; Princeton Holistic Agent Leaderboard slice | no amount publishedno amount published | not dated | party's own page | HAL; Schmidt Sciences | cefb775775 |
| OpenAI; Googleslice · slice | Princeton Holistic Agent Leaderboard slice | no amount publishedAPI and compute credits to evaluate their own models | not dated | party's own page | HAL; HAL paper | 5090d5206e |
| OpenAI; Googleslice · slice | Princeton Holistic Agent Leaderboard slice | no amount publishedno value published (the HAL paper thanked only OpenAI) | not dated | party's own page | HAL; HAL paper | 972615cf46 |
| The Artificial Intelligence Underwriting Company (AIUC; "AI Underwriting Company" on AVERI's list)slice | AVERI slice | no amount publishedListed among AVERI's funders | not dated | party's own page | AVERI, About; AIUC, team | 230b49751e |
| The Artificial Intelligence Underwriting Company (AIUC; "AI Underwriting Company" on AVERI's list)slice | AVERI slice | no amount publishedno amount published | not dated | party's own page | AVERI, About; AIUC, team | af2d5cd33e |
| The Artificial Intelligence Underwriting Company (AIUC; "AI Underwriting Company" on AVERI's list)slice | AVERI slice | figure in prose, not a stated paymentAIUC has raised $55 million from Nat Friedman, Ribbit Capital, First Harmonic and Terrainfrom Nat Friedman, Ribbit Capital, First Harmonic and Terrain | not dated | party's own page | AVERI, About; AIUC, team | 99e0296a2f |
| Halcyon Futuresslice | Transluce slice | no amount publishedTransluce: "Made the Founding Grant" and "Joined Transluce's Board" (Halcyon's testimonial label; no amount) | not dated | party's own page | Halcyon Futures; AVERI, about; ProPublica, Halcyon Futures Inc, Schedule I | a11cd31565 |
| Halcyon Futuresslice | AVERI slice | $611,455$611,455 in grants paid in all | not dated | party's own page | Halcyon Futures; AVERI, about; ProPublica, Halcyon Futures Inc, Schedule I | 9ebe857ffc |
| Halcyon Futuresslice | Transluce slice | no amount publishedTransluce and AVERI are not among the three organization grantees on Schedule I | not dated | party's own page | Halcyon Futures; AVERI, about; ProPublica, Halcyon Futures Inc, Schedule I | 73385ad554 |
| Fathomslice | AVERI slice | no amount publishedNamed among its funders (no amount) | not dated | party's own page | AVERI, about; ProPublica, Halcyon Schedule R | 42b0b0a63e |
| Fathomslice | AVERI slice | no amount publishedlisted as a related tax-exempt organization on Halcyon Futures Inc's FY2024 Schedule R | not dated | party's own page | AVERI, about; ProPublica, Halcyon Schedule R | 4e91df819c |
| Sentinel Bioslice | SecureBio slice | no amount published"We are primarily funded by Coefficient Giving, the Survival and Flourishing Fund, and Sentinel Bio" (no amount) | not dated | 990 | ProPublica, Sentinel Bio FY2024 Form 990; ProPublica, Ergo Impact 2025 Form 990; SecureBio; ProPublica, Halcyon Schedule R | 974fc4e86c |
| Sentinel Bioslice | SecureBio slice | no amount publishedSentinel Bio is a related organization on Halcyon Futures Inc's FY2024 Schedule R and on Ergo Impact Inc's 2025 Schedule R, while Sentinel Bio's own return answers No to related organizations (Part IV, line 34) | 2025 | 990 | ProPublica, Sentinel Bio FY2024 Form 990; ProPublica, Ergo Impact 2025 Form 990; SecureBio; ProPublica, Halcyon Schedule R | 3c0f35751d |
| Sentinel Bioslice | SecureBio slice | $600,000$600,000 (June 2025) to expand AI x Bio risk evaluation from LLMs into biological design tools and build policy capacityto expand AI x Bio risk evaluation from LLMs into biological design tools and build policy capacity | June 2025 | party's own page | Sentinel Bio; Sentinel Bio, 2025 Annual Letter | 0db439f94f |
| Sentinel Bioslice | SecureBio slice | no amount publishedthe Annual Letter adds 'SecureBio's Virology Capabilities Test, which Sentinel funded' | not dated | party's own page | Sentinel Bio; Sentinel Bio, 2025 Annual Letter | 66eedcdfdd |
| Sentinel Bioslice | Epoch AI slice | $80,000$80,000 (Aug 2024), | Aug 2024 | party's own page | Sentinel Bio, Research & Projects; Epoch AI; Sentinel Bio, RAND | 0ca4edb7b2 |
| Sentinel Bioslice | Epoch AI slice | $85,000$85,000 (Sep 2025), | Sep 2025 | party's own page | Sentinel Bio, Research & Projects; Epoch AI; Sentinel Bio, RAND | 4eee235c5d |
| Sentinel Bioslice | Epoch AI slice | $15,000$15,000 (Mar 2026) on Sentinel's list, and Epoch's own page lists | Mar 2026 | party's own page | Sentinel Bio, Research & Projects; Epoch AI; Sentinel Bio, RAND | cb7f8d0445 |
| Sentinel Bioslice | Epoch AI slice | restated$85,000 (Sep 2025) | Sep 2025 | party's own page | Sentinel Bio, Research & Projects; Epoch AI; Sentinel Bio, RAND | 5b2443a2b0 |
| Sentinel Bioslice | Epoch AI slice | restated$80,000 (Sep 2024) | Sep 2024 | party's own page | Sentinel Bio, Research & Projects; Epoch AI; Sentinel Bio, RAND | 17c9992e24 |
| Sentinel Bioslice | RAND CORPORATION | $445,000$445,000 (Nov 2025 detail page; the list shows $435,000) | Nov 2025 | party's own page | Sentinel Bio, Research & Projects; Epoch AI; Sentinel Bio, RAND | 19ba57f3b1 |
| Sentinel Bioslice | RAND CORPORATION | $756,000$756,000 (May 2026), each 'to be spent at the discretion of RAND president and CEO Jason Matheny' | May 2026 | party's own page | Sentinel Bio, Research & Projects; Epoch AI; Sentinel Bio, RAND | bf7fcb06dd |
| Sentinel Bio (incubator)slice | BioTrust | $750,000in kind, stated$750,000 committed (May 2026) 'to incubate and launch a third-party trust infrastructure for managed access to bio-capable AI models'committed 'to incubate and launch a third-party trust infrastructure for managed access to bio-capable AI models' | May 2026 | party's own page | Sentinel Bio, BioTrust | 4ef14ba0d7 |
| Sentinel Bio (incubator)slice | BioTrust | no amount publishedled by two Entrepreneurs-in-Residence at Sentinel Bio | not dated | party's own page | Sentinel Bio, BioTrust | 6e8540b004 |
| Marius Hobbhahn (as Manifund regrantor)slice | GENERALITY LABS slice | no amount publishedSupported the 2025 bridge round | 2025 | grant database | Manifund | 06645e272e |
| UK Department for Science, Innovation and Technologyslice | Trail of Bits slice | £255,600£255,600"AI Safety Guardrails Evaluation"; "The services can only be supplied by Trail of Bits, Inc. owing to technical reasons." | January 2024 | contract notice | Find a Tender | bb82cdb61a |
| UK Department for Science, Innovation and Technologyslice | Gryphon Scientific slice | £267,618£267,618, varied to £392,486.78 (March 2024)Analysis and research for the institute's AI safety programmes | March 2024 | contract notice | Find a Tender; Find a Tender, variation | 1244809986 |
| UK Department for Science, Innovation and Technologyslice | Irregular slice | £195,434£195,434Cyber evaluation and advisory work in the AI Safety Institute | January 2024 | contract notice | Contracts Finder | 0233f38985 |
| UK Department for Science, Innovation and Technologyslice | Gray Swan slice | £128,770£128,770"AI Research - Robust Safeguards and Offensive Cyber Capability Measurement provided by GraySwan AI" | January 2024 | contract notice | Contracts Finder | a7ed3ed53f |
| European AI Officeslice | FAR.AI slice; SecureBio slice | €1,434,080EUR 1,434,080Lot 1, CBRN risk modelling and evaluation; 36 months | December 2025 | contract notice | TED | fd401683c2 |
| European AI Officeslice | Capgemini Consulting, Wavestone, Laboratoire national de metrologie et d'essais | €1,700,000EUR 1,700,000Lot 2, cyber offence risk modelling and evaluation | December 2025 | contract notice | TED | ebda9b46c4 |
| European AI Officeslice | METR slice; Epoch AI slice | €1,167,484EUR 1,167,484Lot 3, loss of control risk modelling and evaluation | December 2025 | contract notice | TED | 421931534e |
| European AI Officeslice | EquiStamp, Clarity AI Research, CeSIA | €881,345EUR 881,345Lot 4, harmful manipulation risk modelling and evaluation; FAR.AI says it red-teams as a subcontractor | December 2025 | contract notice | TED; FAR.AI | ec11f529d0 |
| European AI Officeslice | Technopolis, WIK-Consult, TNO | €1,221,588EUR 1,221,587.50Lot 5, sociotechnical risk modelling and evaluation | December 2025 | contract notice | TED | f0732e1aed |
| European AI Officeslice | A French budget ministry, Laboratoire national de metrologie et d'essais | €968,521EUR 968,521Lot 6, agentic evaluation interface; 24 months | December 2025 | contract notice | TED | 08922b9872 |
| UK Department for Science, Innovation and Technologyslice | Faculty slice | £487,325£487,325 (September 2023)"AI Safety Research Services - Taskforce Faculty" in September 2023 | September 2023 | contract notice | Find a Tender | 48a770001f |
| UK Department for Science, Innovation and Technologyslice | Faculty slice | £1,054,475£1,054,475 (April 2024)Institute research on multi-modal models, jailbreaking, bio data obfuscation and spear-phishing from January to April 2024 | April 2024 | contract notice | Contracts Finder | 7908e73689 |
| UK Department for Science, Innovation and Technologyslice | Faculty slice | £298,630£298,630 (November 2024)the Institute's safeguard surveys partner from November 2024 | November 2024 | contract notice | Contracts Finder | af1712ec09 |
Amount stated 136 of 223Dated 147 of 223Purpose stated 74 of 223Funder profiled 201 of 223Recipient on the roster 161 of 223
Who sits where, and who moved as of 2026-10-17
A seat is one role at one organization, read off a team page, a governance page, an appointment notice or a filing, and sorted by which side of the arrangement it is on: evaluator, lab, government, funder. People are ranked by how many sides they hold at once. The role and organization columns are what the record's own words yield; a seat whose record gives no dates says so. A name adds that person to the slice; a side chip switches that side off.
| Person | Side | Seat, as the record states it | Role | Organization | Dates | Source | Id |
|---|---|---|---|---|---|---|---|
| Beth Barnesslice | evaluator | Founder and CEO, METR (Dec 2023); led ARC Evals from 2022 | Founder and CEO | METRalso names ARC | Dec 2023 – current, as read | METR; METR, 19 Sep 2023; AXRP 34 | e92f2f3f06 |
| Beth Barnesslice | lab | OpenAI, left summer 2022; "I sold all my equity when I was able to" | not statedformer | OPENAIOpenAI | – 2022 | METR; METR, 19 Sep 2023; AXRP 34 | 32dd7f69ab |
| Beth Barnesslice | funder | Pledged 80 percent of her OpenAI equity to Founders Pledge | not stated | not statedalso names OPENAIOpenAI | dates not stated | AXRP 34 | e3f6b8cc2f |
| Ryan Kiddslice | evaluator | Advisor, Halcyon Futures | Advisor | HALCYON FUTURES | dates not stated | ryankidd.ai; MATS, Team | d067bacc3c |
| Ryan Kiddslice | evaluator | co-founder and board member, London Initiative for Safe AI | co-founder and board member | London Initiative for Safe AIno profile | dates not stated | ryankidd.ai; MATS, Team | 5e55538d22 |
| Ryan Kiddslice | evaluator | advisor to Catalyze Impact, Pivotal Research, AI Safety ANZ and BASE; his own site also lists a Catalyze Impact board seat | advisor | Catalyze Impact, Pivotal Research, AI Safety ANZ and BASEno profile | dates not stated | ryankidd.ai; MATS, Team | ff245d481f |
| Ryan Kiddslice | funder | Manifund regrantor | regrantor | Manifundno profile | dates not stated | MATS, Team | da2812b637 |
| Josh Jacobsonslice | evaluator | Formerly METR, FAR AI, Anthropic and the Berkeley Existential Risk Initiative, per MATS's team page (its bio reads "Anthropic, on a research grant") | not statedformer | METRalso names FAR.AI, ANTHROPIC, MATS RESEARCHAnthropic | dates not stated | FAR.AI; MATS, Team; ProPublica | 2711d147d3 |
| Josh Jacobsonslice | evaluator | FAR.AI's page: "interim COO at FAR.AI in 2022 and early 2023", then a board advisor | interim COO | FAR.AI | 2022 – 2023 | FAR.AI | 2ba9cfc9e8 |
| Josh Jacobsonslice | funder | COO, Future of Life Foundation | COO | Future of Life Foundationno profile | dates not stated | MATS, Team; ProPublica | 750631cff8 |
| Michael Airdslice | government | Affiliate, Centre for the Governance of AI | Affiliate | GOVAI (CENTRE for the GOVERNANCE of AI) | dates not stated | MATS, Team; MATS, AI Safety Talent Needs in 2026 | 429437cb9b |
| Michael Airdslice | government | adviser and cofounder, Institute for AI Policy and Strategy | adviser and cofounder | Institute for AI Policy and Strategyno profile | dates not stated | MATS, Team; MATS, AI Safety Talent Needs in 2026 | f5d9cf8edd |
| Michael Airdslice | funder | AI Program Director, Longview Philanthropy, which MATS thanks among the donors without whom it 'would be unable to run upcoming programs' | AI Program Director | LONGVIEW PHILANTHROPYalso names MATS RESEARCH | dates not stated | MATS, Team; MATS, AI Safety Talent Needs in 2026 | 2d7a4ed53f |
| Rumman Chowdhuryslice | evaluator | Founder and CEO, Humane Intelligence PBC | Founder and CEO | Humane Intelligence PBCno profile | dates not stated | Rumman Chowdhury, About; Humane Intelligence, About; Chowdhury, written testimony | 94c1ae8062 |
| Rumman Chowdhuryslice | evaluator | co-founder and Distinguished Advisor, Humane Intelligence nonprofit (CEO until 1 August 2025) | co-founder and Distinguished Advisorformer | Humane Intelligence | – 1 August 2025 | Rumman Chowdhury, About; Humane Intelligence, About | 908f4227d7 |
| Rumman Chowdhuryslice | government | Former U.S. Science Envoy for AI | U.S. Science Envoy for AIformer | not stated | dates not stated | Rumman Chowdhury, About; Humane Intelligence, About; Chowdhury, written testimony | 6f5f251650 |
| Rumman Chowdhuryslice | government | written testimony to the House Science Committee | written testimony | House Science Committeeno profile | dates not stated | Chowdhury, written testimony | 1482ccb0af |
| Rumman Chowdhuryslice | funder | Responsible AI Fellow, Berkman Klein Center | Responsible AI Fellow | Berkman Klein Centerno profile | dates not stated | Rumman Chowdhury, About | 12db348ef3 |
| Rajiv Dattanislice | evaluator | Advisor and Board Member, METR | Advisor and Board Member | METR | dates not stated | METR; ProPublica; AVERI; AIUC | b68035880c |
| Rajiv Dattanislice | evaluator | former COO and Treasurer, METR (FY2024 Form 990) | COO and Treasurerformer | METR | dates not stated | METR; ProPublica; AVERI; AIUC | 22f8fc9732 |
| Rajiv Dattanislice | evaluator | former Founding Advisor, Transluce (per METR); listed author of AVERI's frontier AI auditing paper | Founding Advisorformer | Translucealso names METR | dates not stated | METR; ProPublica; AVERI; AIUC | a5f506406f |
| Rajiv Dattanislice | funder | Co-founder, Artificial Intelligence Underwriting Company (AIUC); "Career Transition Grant" and "Invested in Every Round" on Halcyon Futures's page | Co-founder | AIUCalso names HALCYON FUTURES | dates not stated | METR; ProPublica; AVERI; AIUC; Halcyon Futures | 84a9523aac |
| Chris Painterslice | evaluator | President, METR; earlier Head of Policy | President | METR | dates not stated | METR; testimony; CNAS | 77fa785320 |
| Chris Painterslice | government | Former Technology and National Security Fellow, DoD Joint Artificial Intelligence Center | Technology and National Security Fellowformer | DoD Joint Artificial Intelligence Centerno profile | dates not stated | METR | 1d64a21608 |
| Chris Painterslice | government | Senate HSGAC witness, 30 Sep 2026 | witness | Senate HSGACno profile | 30 Sep 2026 – current, as read | testimony | 4f0ed18740 |
| Chris Painterslice | funder | Adjunct Senior Fellow, CNAS | Adjunct Senior Fellow | CNASno profile | dates not stated | CNAS | 977fea76c1 |
| Ajeya Cotraslice | evaluator | Technical Staff, METR (by Feb 2026) | Technical Staff | METR | Feb 2026 (by) – current, as read | METR; METR; Epoch AI; ProPublica | bf0a7d714f |
| Ajeya Cotraslice | evaluator | report lead, Frontier Risk Report | report lead | Frontier Risk Reportno profile | dates not stated | METR; METR | 23b1a9bc75 |
| Ajeya Cotraslice | evaluator | Board of Directors, Epoch AI (from its early-2025 spin-out) | Board of Directors | Epoch AI | 2025 – current, as read | Epoch AI | 8f41b9b17e |
| Ajeya Cotraslice | evaluator | Director, Redwood Research, to Aug 2024, unpaid | Directorformer | Redwood Research | – Aug 2024 | ProPublica | ce33c050e2 |
| Ajeya Cotraslice | funder | Senior Advisor, Coefficient Giving; managed technical AI safety grantmaking to 2024; "This RFP text was largely drafted by Ajeya Cotra" of the benchmarks RFP that closed "after funding $25 million in grants" | Senior Advisorformer | OPEN PHILANTHROPY (COEFFICIENT GIVING) | – 2024 | METR; METR; Epoch AI; ProPublica; Coefficient Giving; Coefficient Giving | 378fba47ef |
| Ryan Greenblattslice | evaluator | Technical Staff, METR, from 27 Sep 2026 | Technical Staff | METR | 27 Sep 2026 – current, as read | METR; METR; Redwood; arXiv 2412.14093 | dd5a9a03f5 |
| Ryan Greenblattslice | evaluator | METR contractor on the Hugging Face investigation, Jul to Aug 2026 | contractor on the Hugging Face investigationformer | METRalso names HUGGING FACE | – Aug 2026 | METR; METR; Redwood; arXiv 2412.14093 | 8e5443ef6a |
| Ryan Greenblattslice | evaluator | Chief Scientist, Redwood Research, to Sep 2026 | Chief Scientistformer | Redwood Research | – Sep 2026 | METR; METR; Redwood | cc4e186cc2 |
| Ryan Greenblattslice | lab | Co-author with Anthropic's researchers of the alignment-faking paper | Co-author | ANTHROPICAnthropic | dates not stated | arXiv 2412.14093 | c04ea2ffad |
| Buck Shlegerisslice | evaluator | CEO and board member, Redwood Research | CEO and board member | Redwood Research | dates not stated | Redwood | 924f2e4df2 |
| Buck Shlegerisslice | evaluator | board member, Alignment Research Center | board member | ARC | dates not stated | ARC | ede863b9d0 |
| Buck Shlegerisslice | evaluator | board member, Apart Research | board member | Apart Research | dates not stated | Apart | 3bb539aedb |
| Buck Shlegerisslice | lab | Redwood advises Google DeepMind and Anthropic (organization-level) | not stated | Redwood Researchalso names GOOGLE DEEPMIND, ANTHROPICAnthropic · Google DeepMind | dates not stated | Redwood | 1d7ccb5549 |
| Buck Shlegerisslice | government | Expert advisor, UK AISI Alignment Project | Expert advisor | UK AISI | dates not stated | Alignment Project | b91d698b8a |
| Paul Christianoslice | evaluator | Founder and executive director, Alignment Research Center, from which METR spun out | Founder and executive director | ARCalso names METR | dates not stated | ARC; METR, 19 Sep 2023 | 3b139b8428 |
| Paul Christianoslice | lab | OpenAI alignment research, 2017 to 2021 | alignment researchformer | OPENAIOpenAI | 2017 – 2021 | ARC; METR, 19 Sep 2023; OpenAI Foundation; Anthropic | 66ad162396 |
| Paul Christianoslice | lab | OpenAI Foundation board and Safety and Security Committee, 9 Sep 2026 | board and Safety and Security Committee | OPENAI FOUNDATIONOpenAI | 9 Sep 2026 – current, as read | ARC; METR, 19 Sep 2023; OpenAI Foundation; Anthropic | 234c3a9a77 |
| Paul Christianoslice | lab | initial trustee of Anthropic's Long-Term Benefit Trust, stepped down Apr 2024 | initial trusteeformer | ANTHROPICAnthropic | – Apr 2024 | OpenAI Foundation; Anthropic | 49a83d95b2 |
| Paul Christianoslice | government | Head of AI Safety, US AI Safety Institute, from Apr 2024 | Head of AI Safety | US CAISI | Apr 2024 – current, as read | ARC; METR, 19 Sep 2023; OpenAI Foundation; Anthropic | 193ffc692c |
| Paul Christianoslice | government | Senior Technical Advisor, CAISI, recused from OpenAI matters and model evaluations | Senior Technical Advisor | US CAISIalso names OPENAIOpenAI | dates not stated | ARC; METR, 19 Sep 2023; OpenAI Foundation; Anthropic | 7103ade165 |
| Paul Christianoslice | government | UK Frontier AI Taskforce advisory board, Sep 2023 | advisory board | UK AISI | Sep 2023 – current, as read | NIST; GOV.UK | 92952df282 |
| Marius Hobbhahnslice | evaluator | Co-founder and CEO, Apollo Research | Co-founder and CEO | Apollo Research | dates not stated | Hobbhahn CV; European Commission; testimony; Manifund | cbba38f16f |
| Marius Hobbhahnslice | evaluator | part-time research fellow, Epoch, 2022 to 2023 | part-time research fellowformer | Epoch AI | 2022 – 2023 | Hobbhahn CV | ebf0e84d58 |
| Marius Hobbhahnslice | government | Member, European Commission Scientific Panel of Independent Experts on AI | Member | European Commission Scientific Panel of Independent Experts on AIno profile | dates not stated | European Commission | d47e9bd62c |
| Marius Hobbhahnslice | government | Senate HSGAC witness, 30 Sep 2026 | witness | Senate HSGACno profile | 30 Sep 2026 – current, as read | testimony | d744703cb6 |
| Marius Hobbhahnslice | funder | Manifund regrantor who supported the 2025 bridge round of the Inspect Evals project now held by Generality Labs | Manifund regrantor who supported the bridge round | GENERALITY LABS | 2025 – current, as read | Manifund | 9ac079fd08 |
| Dan Lahavslice | evaluator | Co-founder and CEO, Irregular; holds equity | Co-founder and CEO | Irregular | dates not stated | paper; arXiv; RAND | 8621213fac |
| Dan Lahavslice | government | Co-author with EU AI Office and Joint Research Centre staff | Co-author | EU AI OFFICE | dates not stated | arXiv | 37a40f17cf |
| Dan Lahavslice | funder | Co-author of RAND's "Securing AI Model Weights" | Co-author | RAND CORPORATION | dates not stated | RAND | 9adb10ac08 |
| Omer Nevoslice | evaluator | Co-founder and CTO, Irregular | Co-founder and CTO | Irregular | dates not stated | Globes | 6b6bc5178e |
| Omer Nevoslice | lab | Formerly Google Research development manager (reported) | development managerformer | Google Researchno profileGoogle DeepMind | dates not stated | Calcalist | e51fef0170 |
| Omer Nevoslice | government | Unit 8200, twelve years (reported) | not stated | Unit 8200no profile | dates not stated | Calcalist | 21f8de9a61 |
| Omer Nevoslice | funder | Co-founder and board member, Effective Altruism Israel, volunteer | Co-founder and board member | Effective Altruism Israelno profile | dates not stated | 687a4a5813 | |
| Matt Fredriksonslice | evaluator | Co-founder and CEO, Gray Swan | Co-founder and CEO | Gray Swan | dates not stated | Gray Swan | f47ecc7f32 |
| Matt Fredriksonslice | government | Associate Professor, Carnegie Mellon | Associate Professor | Carnegie Mellonno profile | dates not stated | Carnegie Mellon | 7c42bf4d9f |
| Zico Kolterslice | evaluator | Co-founder and Chief Scientist, Gray Swan | Co-founder and Chief Scientist | Gray Swan | dates not stated | Gray Swan | 1c756254b7 |
| Zico Kolterslice | lab | OpenAI board from 8 Aug 2024 | board | OPENAIOpenAI | 8 Aug 2024 – current, as read | Gray Swan; OpenAI; OpenAI; Alignment Project | 76e6328bdb |
| Zico Kolterslice | lab | chair of its Safety and Security Committee from 16 Sep 2024 | chair | OPENAIOpenAI | 16 Sep 2024 – current, as read | Gray Swan; OpenAI; OpenAI; Alignment Project | 37961ebe60 |
| Zico Kolterslice | lab | OpenAI Foundation board and observer on the OpenAI Group board after Oct 2025 | OpenAI Foundation board and observer | OPENAIalso names OPENAI FOUNDATIONOpenAI | Oct 2025 – current, as read | OpenAI; OpenAI | fb571cc3df |
| Zico Kolterslice | government | Expert advisor, UK AISI Alignment Project | Expert advisor | UK AISI | dates not stated | Alignment Project | 87eadae2e6 |
| Zico Kolterslice | government | Professor and department head, Carnegie Mellon | Professor and department head | Carnegie Mellonno profile | dates not stated | Gray Swan; OpenAI; OpenAI; Alignment Project | d415107cfd |
| Zico Kolterslice | government | Qualcomm board | board | Qualcommno profile | dates not stated | SEC | 6b7541e800 |
| Dan Hendrycksslice | evaluator | Adviser to Gray Swan; announced as Chief Strategy Advisor 16 Jul 2024, divested his equity 25 Jul 2024 and stayed "as an advisor, without pay" | Adviser | Gray Swan | dates not stated | his post; arXiv 2504.16137; Fortune | ee2cfd2162 |
| Dan Hendrycksslice | evaluator | co-author of SecureBio's Virology Capabilities Test | co-author | SecureBio | dates not stated | arXiv 2504.16137 | eaac9465bc |
| Dan Hendrycksslice | evaluator | adviser to Scale AI (reported) | adviser | Scale AI | dates not stated | Fortune | f76c28706b |
| Dan Hendrycksslice | lab | Safety adviser, xAI (reported) | Safety adviser | XAI (SPACEXAI)xAI | dates not stated | Fortune | 633d533671 |
| Dan Hendrycksslice | funder | Director, Center for AI Safety | Director | CENTER for AI SAFETY | dates not stated | CAIS | da01348620 |
| Kevin Esveltslice | evaluator | Co-founder of SecureBio; unpaid director on its Form 990s for 2022 to 2024; now listed under past members | Co-founderformer | SecureBio | – 2024 | Substack; ProPublica; SecureBio; New York Times, via MIT; S. Hrg. 118-376 | 4e2aa9ff9a |
| Kevin Esveltslice | lab | "has consulted for Anthropic and OpenAI" (reported) | has consulted | ANTHROPIC; OPENAIOpenAI · Anthropic | dates not stated | New York Times, via MIT | 7f8a02b950 |
| Kevin Esveltslice | government | Senate HSGAC witness, 11 Jul 2024 | witness | Senate HSGACno profile | 11 Jul 2024 – current, as read | S. Hrg. 118-376 | a194d6908e |
| Jaime Sevillaslice | evaluator | Co-founder, CEO and board member, Epoch AI; Treasurer on its FY2024 990-EZ | Co-founder, CEO and board member | Epoch AI | dates not stated | Epoch AI; ProPublica; jaimesevilla.me | dcbd014ebf |
| Jaime Sevillaslice | funder | Advisor, Observatorio de Riesgos Catastróficos Globales | Advisor | Observatorio de Riesgos Catastróficos Globalesno profile | dates not stated | jaimesevilla.me | d75aabab8a |
| Tamay Besirogluslice | evaluator | Co-founder, Epoch AI; associate director to 2025, then research advisor | Co-founderformer | Epoch AI | – 2025 | Epoch AI; Mechanize; his post | 35e2a3cf5e |
| Tamay Besirogluslice | evaluator | co-founder, Mechanize, Apr 2025 | co-founder | MECHANIZE | Apr 2025 – current, as read | Mechanize | a4635dce1e |
| Tamay Besirogluslice | lab | Research scientist, Google DeepMind, by Sep 2026 | Research scientist | GOOGLE DEEPMINDGoogle DeepMind | Sep 2026 (by) – current, as read | his post | 023e982cf3 |
| Ege Erdilslice | evaluator | Researcher, Epoch AI, left by Mar 2025 | Researcherformer | Epoch AI | – Mar 2025 | Epoch AI; Mechanize | 12a61628bc |
| Ege Erdilslice | evaluator | co-founder, Mechanize | co-founder | MECHANIZE | dates not stated | Epoch AI; Mechanize | ad59928401 |
| Adam Gleaveslice | evaluator | Co-founder and CEO, FAR.AI; "Advisor and Board Member," METR | Co-founder and CEO | FAR.AIalso names METR | dates not stated | METR; Adam Gleave; European Commission | bc848e2fb8 |
| Adam Gleaveslice | evaluator | board, Safe AI Forum and London Initiative for Safe AI | board | Safe AI Forum and London Initiative for Safe AIno profile | dates not stated | Adam Gleave | 27efd8b0b7 |
| Adam Gleaveslice | lab | Formerly Google DeepMind | not statedformer | GOOGLE DEEPMINDGoogle DeepMind | dates not stated | Adam Gleave | 88454ad667 |
| Adam Gleaveslice | government | Member, European Commission AI Act Scientific Panel, from Jun 2026 | Member | European Commission AI Act Scientific Panelno profile | Jun 2026 – current, as read | European Commission | 0c9eb08d45 |
| Adam Gleaveslice | funder | Schmidt Sciences AI2050 Early Career Fellow, Nov 2025 | AI2050 Early Career Fellow | SCHMIDT SCIENCES | Nov 2025 – current, as read | FAR.AI | fa3390c3fc |
| Jeffrey Ladishslice | evaluator | President and Executive Director, Palisade Research | President and Executive Director | Palisade Research | dates not stated | Palisade | d8ee513fdf |
| Jeffrey Ladishslice | lab | "helped build out the information security program at Anthropic"; left in 2022 | helped build out the information security programformer | ANTHROPICAnthropic | – 2022 | Palisade; Foresight Institute | e3d44fd61a |
| Jeffrey Ladishslice | government | "has helped advise the White House, Department of Defense, and congressional offices" | not stated | not stated | dates not stated | Palisade | de7ac09309 |
| Jacob Steinhardtslice | evaluator | Co-founder, CEO and director, Transluce | Co-founder, CEO and director | Transluce | dates not stated | Transluce | 69ed8a65b7 |
| Jacob Steinhardtslice | lab | Post-doc year at OpenAI and Open Philanthropy | Post-doc year | OPENAI; OPEN PHILANTHROPY (COEFFICIENT GIVING)OpenAI | dates not stated | Berkeley | 2220b13984 |
| Jacob Steinhardtslice | government | Associate Professor, UC Berkeley | Associate Professor | UC Berkeleyno profile | dates not stated | Transluce; Berkeley; SCAI 2025; LawZero | 753d0b204c |
| Jacob Steinhardtslice | government | CSET fellow | fellow | CSETno profile | dates not stated | SCAI 2025 | e2b3eb4672 |
| Jacob Steinhardtslice | funder | Scientific Advisor, LawZero | Scientific Advisor | LAWZERO | dates not stated | LawZero | 997d8fd4ce |
| Sarah Schwettmannslice | evaluator | Co-founder and Chief Scientist, Transluce; officer at $184,273 on the fiscal 2024 Form 990 | Co-founder and Chief Scientist | Transluce | dates not stated | Transluce; ProPublica; her site | 803b39329e |
| Sarah Schwettmannslice | government | Research Scientist, MIT CSAIL | Research Scientist | MIT CSAILno profile | dates not stated | her site | 185a979336 |
| Geoffrey Irvingslice | evaluator | Co-founder and Chief Scientist, Resolution, Jun 2026 | Co-founder and Chief Scientist | RESOLUTION | Jun 2026 – current, as read | naml.us | 868acbbc75 |
| Geoffrey Irvingslice | lab | OpenAI (Reflection Team lead) and Google DeepMind (Scalable Alignment lead), before 2024 | Reflection Team lead | OPENAIalso names GOOGLE DEEPMINDOpenAI · Google DeepMind | 2024 – current, as read | UK AISI | 519a3e630d |
| Geoffrey Irvingslice | government | Research Director, then Chief Scientist, UK AI Security Institute, Apr 2024 to May 2026; advisory capacity after | Research Director, then Chief Scientistformer | UK AISI | Apr 2024 – May 2026 | naml.us; UK AISI; naml.us; LawZero | cc2868a922 |
| Geoffrey Irvingslice | funder | Scientific Advisor, LawZero | Scientific Advisor | LAWZERO | dates not stated | LawZero | c4e84ff666 |
| Ian Hogarthslice | lab | Sold his Anthropic seed stake, with Conjecture, Lakera and Faculty, on appointment | not stated | not statedalso names ANTHROPIC, FacultyAnthropic | dates not stated | GOV.UK register | f425278278 |
| Ian Hogarthslice | government | Chair, UK AI Security Institute (Foundation Model Taskforce from 12 Jun 2023) | Chair | UK AISI | 12 Jun 2023 – current, as read | GOV.UK | 5005a935b2 |
| Ian Hogarthslice | funder | General Partner, Plural | General Partner | Pluralno profile | dates not stated | GOV.UK register | 079639d691 |
| Jade Leungslice | lab | Governance Lead, OpenAI, to Oct 2023 | Governance Leadformer | OPENAIOpenAI | – Oct 2023 | GovAI | 80fb9a69d9 |
| Jade Leungslice | government | CTO, UK AI Safety Institute (Oct 2023; CTO by May 2024) | CTO | UK AISI | Oct 2023 – current, as read | GovAI; AISI; GOV.UK | 1089f4d888 |
| Jade Leungslice | government | Prime Minister's Adviser on AI, 15 Aug 2025 | Prime Minister's Adviser on AI | not stated | 15 Aug 2025 – current, as read | GovAI; AISI; GOV.UK | 6f156c798e |
| Jade Leungslice | government | Vice-Chair of the Institute, 23 Sep 2026 | Vice-Chair of the Institute | not stated | 23 Sep 2026 – current, as read | AISI; GOV.UK | 3a3dfaa6f8 |
| Jade Leungslice | funder | Co-founded GovAI as Head of Research and Partnerships | co-founder; Head of Research and Partnerships | GOVAI (CENTRE for the GOVERNANCE of AI) | dates not stated | GovAI | 03cedd71d0 |
| Dan Guidoslice | evaluator | CEO and co-founder, Trail of Bits | CEO and co-founder | Trail of Bits | dates not stated | Trail of Bits | 0ba770089f |
| Dan Guidoslice | lab | OpenAI: Patch the Planet partner; reported acquisition talks, Oct 2026 | Patch the Planet partner | OPENAIOpenAI | dates not stated | Trail of Bits; OpenAI; PYMNTS; CFTC | af0531c2af |
| Dan Guidoslice | government | Co-chair, Cybersecurity Subcommittee, CFTC Technology Advisory Committee, 2023 | Co-chair | Cybersecurity Subcommittee, CFTC Technology Advisory Committeeno profile | 2023 – current, as read | CFTC | 2b4c3dadaf |
| Lukas Peterssonslice | evaluator | Co-founder and CEO, Andon Labs | Co-founder and CEO | Andon Labs | dates not stated | his CV | 8e37ca26e1 |
| Lukas Peterssonslice | lab | Andon built the infrastructure of Anthropic's Project Vend (organization-level) | not stated | Andon Labsalso names ANTHROPICAnthropic | dates not stated | Anthropic | 7ceb1e22bd |
| Helen Tonerslice | lab | OpenAI board, 8 Sep 2021 to Nov 2023 | boardformer | OPENAIOpenAI | 8 Sep 2021 – Nov 2023 | OpenAI; OpenAI | 712c62d385 |
| Helen Tonerslice | government | Interim Executive Director, Georgetown CSET, from 2 Sep 2025 | Interim Executive Director | Georgetown CSETno profile | 2 Sep 2025 – current, as read | CSET | 0abfa40ec8 |
| Helen Tonerslice | funder | Former senior research analyst, Open Philanthropy | senior research analystformer | OPEN PHILANTHROPY (COEFFICIENT GIVING) | dates not stated | CSET | ab5dc0bd38 |
| Holden Karnofskyslice | evaluator | Board, Alignment Research Center, to Apr 2023 | Boardformer | ARC | – Apr 2023 | ProPublica | 0f3e9faea4 |
| Holden Karnofskyslice | lab | Member of technical staff, Anthropic, from Jan 2025 (reported) | Member of technical staff | ANTHROPICAnthropic | Jan 2025 – current, as read | Fortune | 4baecf5f55 |
| Holden Karnofskyslice | lab | OpenAI board, 2017 to 2021 (reported) | boardformer | OPENAIOpenAI | 2017 – 2021 | TechCrunch | 932fbce2fc |
| Holden Karnofskyslice | funder | Co-founder and former CEO and co-CEO, Open Philanthropy, to 29 Apr 2024; co-founder, GiveWell | Co-founder and former CEO and co-CEOformer | OPEN PHILANTHROPY (COEFFICIENT GIVING) | – 29 Apr 2024 | EA Forum; GiveWell | 8647c5a05d |
| Kyle Scottslice | between | Board member, Palisade Research, "previously the Operations Manager at" METR | Board member | Palisade Researchalso names METR | dates not stated | Palisade; ARC | 1acbe24f08 |
| Kyle Scottslice | between | Treasurer, Alignment Research Center | Treasurer | ARC | dates not stated | Palisade; ARC | f4ef777b22 |
| Sawyer Bernathslice | between | Board member, FAR.AI, and board member, SecureBio | Board member | FAR.AIalso names SecureBio | dates not stated | FAR.AI; SecureBio | e1ed743e81 |
| Sawyer Bernathslice | between | Executive Director of the Berkeley Existential Risk Initiative 2019 to 2024 | Executive Directorformer | Berkeley Existential Risk Initiativeno profile | 2019 – 2024 | FAR.AI; SecureBio | cfa6fed72c |
| Sawyer Bernathslice | between | Operations Director, Tarbell Center for AI Journalism | Operations Director | Tarbell Center for AI Journalismno profile | dates not stated | FAR.AI; SecureBio | aaa5174972 |
| Max Hendersonslice | between | Board member, AVERI ("Max is a board member at AVERI and the founder of Ergo Impact") | Board member | AVERIalso names ERGO IMPACT INC | dates not stated | AVERI; Halcyon Futures; ProPublica, Halcyon Schedule R | 174e764983 |
| Max Hendersonslice | between | Venture Partner, Halcyon Futures ("Max is the Co-founder and CEO of Ergo Impact"); the same name is Secretary on Halcyon's and Treasurer on Sentinel Bio's FY2024 Form 990, CEO on Ergo Impact's 2025 return, and a board member on Fathom's | Venture Partner | HALCYON FUTURESalso names ERGO IMPACT INC, SENTINEL BIO, FATHOM | dates not stated | AVERI; Halcyon Futures; ProPublica, Halcyon Schedule R | f482a86c6b |
| Kellie Liketslice | between | Secretary and director, Sentinel Bio Inc (FY2024 Form 990) | Secretary and director | SENTINEL BIO | dates not stated | ProPublica, Sentinel Bio FY2024 Form 990; Sentinel Bio, People; Ergo Impact; ProPublica, Halcyon Schedule R | 284f7581cb |
| Kellie Liketslice | between | co-founder and President, Ergo Impact | co-founder and President | ERGO IMPACT INC | dates not stated | ProPublica, Sentinel Bio FY2024 Form 990; Sentinel Bio, People; Ergo Impact; ProPublica, Halcyon Schedule R | 70b0153dc9 |
| Kellie Liketslice | between | Director, Halcyon Futures Inc | Director | HALCYON FUTURES | dates not stated | ProPublica, Sentinel Bio FY2024 Form 990; Sentinel Bio, People; Ergo Impact; ProPublica, Halcyon Schedule R | a29cf402b9 |
| Kellie Liketslice | between | board member, Fathom AI (the same name across the four organizations' records) | board member | FATHOM | dates not stated | ProPublica, Sentinel Bio FY2024 Form 990; Sentinel Bio, People; Ergo Impact; ProPublica, Halcyon Schedule R | 32da4548ae |
| Cari Tunaslice | between | Chair of Good Ventures and of the Boards of Managers/Directors of Coefficient Giving LLC and Coefficient Giving Research | Chair | GOOD VENTURESalso names OPEN PHILANTHROPY (COEFFICIENT GIVING) | dates not stated | Coefficient Giving, Governance; ProPublica, Good Ventures Form 990-PF | 2130e9a164 |
| Cari Tunaslice | between | President and Board Chair on Good Ventures' Form 990-PF | President and Board Chair | GOOD VENTURES | dates not stated | Coefficient Giving, Governance; ProPublica, Good Ventures Form 990-PF | c6d3c9ed26 |
| Grace Wernerslice | between | Board member and co-founder, AVERI ("Previously, Grace co-founded AVERI and served as Chief of Staff and lead U.S. policy") | Board member and co-founder | AVERI | dates not stated | AVERI; FAI; AEF-1 | f5906571bd |
| Grace Wernerslice | between | Vice President of Growth, Foundation for American Innovation | Vice President of Growth | FOUNDATION for AMERICAN INNOVATION | dates not stated | AVERI; FAI; AEF-1 | c38639d19c |
| Grace Wernerslice | between | author of AEF-1 and of the 48-author "Frontier AI Auditing" paper | author | AEF-1no profile | dates not stated | AVERI; FAI; AEF-1 | 2136c1689f |
| Patricia Paskovslice | between | Standards Director, AVERI ("she leads the development of standards for frontier AI auditing") | Standards Director | AVERI | dates not stated | AVERI; AI Evaluator Forum; RAND | 0bffe7ff0a |
| Patricia Paskovslice | between | Senior Advisor, AI Evaluator Forum | Senior Advisor | AI EVALUATOR FORUM | dates not stated | AVERI; AI Evaluator Forum; RAND | 4318e0d929 |
| Patricia Paskovslice | between | adjunct researcher at RAND | adjunct researcher | RAND CORPORATION | dates not stated | AVERI; AI Evaluator Forum; RAND | 6105564947 |
| Patricia Paskovslice | between | Research Affiliate, Oxford Martin AI Governance Initiative; author of AEF-1 | Research Affiliate | Oxford Martin AI Governance Initiativeno profile | dates not stated | AVERI; AI Evaluator Forum; RAND | cb64991b9b |
| Lawrence Chanslice | between | Board member, FAR.AI | Board member | FAR.AI | dates not stated | FAR.AI | 05e5bce0aa |
| Lawrence Chanslice | between | formerly of Redwood Research, per his FAR.AI bio | not statedformer | Redwood Researchalso names FAR.AI | dates not stated | FAR.AI | b81a3dcc95 |
| Bri Treeceslice | between | Board member, FAR.AI; "served as the COO of the Center for AI Safety (CAIS) and founded the Center for AI Safety Action Fund" | Board member | FAR.AI | dates not stated | FAR.AI; Fathom; Fathom, SB 813 | 4f6d402f1a |
| Bri Treeceslice | between | co-founder and President of Fathom, which sponsored California SB 813, the bill that has the state designate independent verification organizations for AI | co-founder and President | FATHOM | dates not stated | FAR.AI; Fathom; Fathom, SB 813 | f4561bfc9e |
| Ben Goldhaberslice | between | Director of FAR.Labs on FAR.AI's 2024 filing; "joined Coefficient Giving in April 2026" | Director | FAR.AIalso names OPEN PHILANTHROPY (COEFFICIENT GIVING) | 2024 – current, as read | FAR.AI Form 990, 2024; Coefficient Giving | 93256a1b1d |
| Daniel Kokotajloslice | between | First mission director on Apollo Research's board; listed research advisor, Epoch AI | First mission director | Apollo Research | dates not stated | Apollo; Epoch AI | fd31f379bf |
| Yoshua Bengioslice | between | Adviser, METR | Adviser | METR | dates not stated | LawZero; METR; Transluce; Alignment Project | 93376e45c3 |
| Yoshua Bengioslice | between | adviser, Transluce | adviser | Transluce | dates not stated | LawZero; METR; Transluce; Alignment Project | e085c4fa4f |
| Yoshua Bengioslice | between | UK AISI Alignment Project advisory board | advisory board | UK AISI | dates not stated | LawZero; METR; Transluce; Alignment Project | 5fa635ef91 |
| Yoshua Bengioslice | between | LawZero, where Steinhardt and Irving are scientific advisors, as is METR's chief executive Beth Barnes | not stated | LAWZEROalso names METR | dates not stated | LawZero | 55144d6194 |
| Marco Mascorroslice | between | Adviser, METR | Adviser | METR | dates not stated | METR; Mechanize | 3e81dfb1d4 |
| Marco Mascorroslice | between | named backer of Mechanize, including its April 2026 round | named backer | MECHANIZE | April 2026 – current, as read | METR; Mechanize | f8cd01d5e7 |
| Percy Liangslice | between | Director, Stanford CRFM | Director | Stanford CRFM | dates not stated | Transluce; Percy Liang | 5aa38f987e |
| Percy Liangslice | between | adviser to Transluce, whose fundraiser carries his endorsement | adviser | Transluce | dates not stated | Transluce; Percy Liang | 28e8f27e23 |
| Conrad Stoszslice | between | Head of Governance, Transluce, from 22 Oct 2025; "previously led the U.S. Center for Standards and Innovation" | Head of Governance | Translucealso names US CAISI | 22 Oct 2025 – current, as read | Transluce | f2aada7bba |
| J.J. Allaireslice | between | Co-founder and Principal Engineer, Meridian Labs; "Technical Staff" at UK AISI in Nov 2024; "a visiting researcher at the US CAISI" | Co-founder and Principal Engineer | MERIDIAN LABSalso names UK AISI | dates not stated | Meridian Labs; UK AISI | a3568f5329 |
| Alexandra Abbasslice | between | Technical Program Manager, Meridian Labs; "co-founded the AI Safety Engineering Taskforce in 2024," the Arcadia Impact programme from which Generality Labs spun out | Technical Program Manager | MERIDIAN LABSalso names GENERALITY LABS | dates not stated | Meridian Labs; Manifund | e6becd0a0d |
| Kai Fronsdalslice | between | Meridian Labs; "was an Anthropic Fellow and MATS AI Safety Researcher, where he developed Petri" | not stated | MERIDIAN LABSalso names ANTHROPIC, MATS RESEARCHAnthropic | dates not stated | Meridian Labs; Meridian Labs | 9731ce8e36 |
| Kai Fronsdalslice | between | UK AISI research affiliate | research affiliate | UK AISI | dates not stated | Meridian Labs; Meridian Labs | 57fc2938d2 |
| Erin Robertsonslice | between | Arcadia Impact founder and co-director, now at Coefficient Giving | founder and co-director | Arcadia Impactno profilealso names OPEN PHILANTHROPY (COEFFICIENT GIVING) | dates not stated | Coefficient Giving | b58a8f59b7 |
| Eli Tyreslice | between | Head of Strategy, Palisade Research; "supports grantmaking at Jaan Tallinn's Survival and Flourishing Fund" | Head of Strategy | Palisade Research | dates not stated | Palisade | 2d03ab36c1 |
| Benjamin Weinstein-Raunslice | between | Head of Research, Palisade Research | Head of Research | Palisade Research | dates not stated | Palisade; SFF 2026 | ae480a8f8f |
| Benjamin Weinstein-Raunslice | between | acting director of AI Impacts per Palisade's page | acting director | Palisade Research | dates not stated | Palisade; SFF 2026 | 2eb44b1851 |
| Benjamin Weinstein-Raunslice | between | formerly technical staff at SecureDNA, Redwood Research and MIRI | technical staff at SecureDNAformer | Redwood Research | dates not stated | Palisade; SFF 2026 | b4dd79b58a |
| Benjamin Weinstein-Raunslice | between | SFF-2026 Main Track recommender, in the round that recommends Palisade $1,842,000 and AI Impacts $351,000 (the page prints track totals and no recusal rule) | recommender | SURVIVAL and FLOURISHING FUNDalso names Palisade Research | 2026 – current, as read | Palisade; SFF 2026 | 4da97e9250 |
| Blake Borgesonslice | between | Board member, Palisade Research; "helped lead the Machine Intelligence Research Institute for four years" and remains on MIRI's board, a director since September 2018 after a $300,000 donation in 2016 | Board member | Palisade Researchalso names MACHINE INTELLIGENCE RESEARCH INSTITUTE | dates not stated | Palisade; MIRI team; Recursion DEF 14A; ProPublica; AVERI, About | 0fef887b8c |
| Blake Borgesonslice | between | co-founder and director, Recursion Pharmaceuticals | co-founder and director | Recursion Pharmaceuticalsno profile | dates not stated | Palisade; MIRI team; Recursion DEF 14A; ProPublica | 104eaad512 |
| Blake Borgesonslice | between | director, Good Forever Foundation (assets $18,113,687, FY2024; "Good Forever Foundation" is also on AVERI's funder list, a name match) | director | Good Forever Foundationno profilealso names AVERI | dates not stated | Palisade; MIRI team; Recursion DEF 14A; ProPublica | ebc66ba0bd |
| Luke Muehlhauserslice | between | Managing Director, Coefficient Giving's AI fund | Managing Director | OPEN PHILANTHROPY (COEFFICIENT GIVING) | dates not stated | Coefficient Giving; Anthropic | a8130009f0 |
| Luke Muehlhauserslice | between | Anthropic board until 29 May 2024 | boardformer | ANTHROPICAnthropic | – 29 May 2024 | Coefficient Giving; Anthropic | 1504fe66de |
| Miles Brundageslice | between | Strategy advisor, Epoch AI | Strategy advisor | Epoch AI | dates not stated | Epoch AI | 7284f229c3 |
| Miles Brundageslice | between | formerly OpenAI | not statedformer | OPENAIOpenAI | dates not stated | Epoch AI | f40bf45f13 |
| Mike McCormickslice | between | Director, Transluce | Director | Transluce | dates not stated | Transluce; Steinhardt | 41e817958c |
| Mike McCormickslice | between | CEO of Halcyon Futures, which Steinhardt says "was instrumental in helping Transluce get off the ground" | CEO | HALCYON FUTURESalso names Transluce | dates not stated | Transluce; Steinhardt | 1e49e7b2cc |
| Alexandr Wangslice | between | Founder and Chairman, Scale AI, the US institute's first authorized outside evaluator | Founder and Chairman | Scale AI | dates not stated | Scale AI; Meta | df2665bcaf |
| Alexandr Wangslice | between | Chief AI Officer, Meta | Chief AI Officer | METAMeta | dates not stated | Scale AI; Meta | 972eca668c |
| Camille Françoisslice | between | President and CEO, ROOST, whose founding partners include OpenAI and Google and whose board seated an OpenAI Chief Strategy Officer and a Google vice president on 5 Oct 2026 | President and CEO | ROOST (ROBUST OPEN ONLINE SAFETY TOOLS)also names OPENAIOpenAI · Google DeepMind | 5 Oct 2026 – current, as read | ROOST; ROOST, 5 Oct 2026 | 01c6f518e4 |
Role read 166 of 179Organization read 173 of 179Organization profiled 136 of 179Dated 54 of 179
Moves 32 rows, as of 2026-10-17
| Person | From | To | When | Source | Id | |
|---|---|---|---|---|---|---|
| Dean Ballslice | Fathom (consulting fellow, Jan 2025); White House OSTP (senior policy advisor for AI, mid-April to 11 August 2025); Foundation for American Innovation (senior fellow)→ FATHOM · → FOUNDATION for AMERICAN INNOVATION | → | OpenAI (Head of Strategic Futures, 6 Jul 2026); remains FAI non-resident senior fellow→ OPENAI · → FOUNDATION for AMERICAN INNOVATION | 2025 to 2026 | Hyperdimensional; MeriTalk; FAI | bd8ae43145 |
| Quentin Feuillade-Montixino profileslice | MATS scholar; METR (contractor, evaluating GPT-4 pre-release)→ MATS RESEARCH · → METR | → | PRISM Eval (co-founder); WeaveMind (Seldon Lab batch 2)no profile | after MATS | MATS, Alumni | 77ef05ac7c |
| Joseph Bloomno profileslice | MATS 5.0 scholar (mentor Neel Nanda)→ MATS RESEARCH | → | UK AI Security Institute (Head of Model Transparency)→ UK AISI | by 2026 | MATS | 9f6bc92b50 |
| Rohan Subramani; Steven Veldno profileslice | MATS program (Subramani also founded and led Aether; Veld also IAPS Fellowship and Constellation Astra Fellowship program management)→ MATS RESEARCH | → | Coefficient Giving (May 2026; Sep 2026)→ OPEN PHILANTHROPY (COEFFICIENT GIVING) | 2026 | Coefficient Giving | e3a10fb527 |
| Mala Kumarslice | MLCommons (Director of Program Management, AI Safety); GitHub (Director of Tech for Social Good)→ MLCOMMONS | → | Humane Intelligence (Head of Impact Mar 2025; Interim Executive Director 1 Aug 2025; Executive Director Apr 2026)→ Humane Intelligence | 2025 to 2026 | Humane Intelligence, Team; Humane Intelligence, About | 21cb4bdf13 |
| Ben Kinsellano profileslice | OpenAI (Human Data team)→ OPENAI | → | Humane Intelligence (President of the Board)→ Humane Intelligence | by 2026 | Humane Intelligence, Board & Advisory | 433df3ac99 |
| Kristi Arbogastno profileslice | Graphika (OSINT and content-moderation investigations into generative AI harms); National Democratic Institute→ GRAPHIKA | → | Humane Intelligence (AI Red Teaming Manager); NIST (AI Governance and Risk Analyst)→ Humane Intelligence | by 2026 | Humane Intelligence, Team | d241b5cf53 |
| Beth Barnesslice | OpenAI→ OPENAI | → | Alignment Research Center (ARC Evals), then METR→ ARC · → METR | Summer 2022; Dec 2023 | AXRP 34; METR | 208503bfc8 |
| Paul Christianoslice | OpenAI (2017 to 2021)→ OPENAI | → | ARC; US AI Safety Institute (Apr 2024); CAISI senior technical adviser; back to ARC as executive director (Aug 2026); OpenAI Foundation board (9 Sep 2026)→ OPENAI FOUNDATION · → ARC · → US CAISI | 2021 to Sep 2026 | NIST; OpenAI Foundation; ARC | 30fdcfdee3 |
| Ajeya Cotraslice | Coefficient Giving , technical AI safety grantmaking; Redwood Research board (to Aug 2024)→ OPEN PHILANTHROPY (COEFFICIENT GIVING) · → Redwood Research | → | METR technical staff; Epoch AI board→ Epoch AI · → METR | 2024 to 2026 | Coefficient Giving; METR; Epoch AI | 8a94650c89 |
| Ryan Greenblattslice | Redwood Research (Chief Scientist)→ Redwood Research | → | METR (contractor Jul to Aug 2026; technical staff 27 Sep 2026)→ METR | 2026 | METR; his post | 9f8348de0c |
| Tamay Besirogluslice | Epoch AI (co-founder, associate director)→ Epoch AI | → | Mechanize (co-founder, Apr 2025); Google DeepMind (research scientist, by Sep 2026)→ MECHANIZE · → GOOGLE DEEPMIND | 2025 to 2026 | Epoch AI; Mechanize; his post | d6d4d94fda |
| Ege Erdilslice | Epoch AI→ Epoch AI | → | Mechanize (co-founder)→ MECHANIZE | Apr 2025 | Mechanize | 8b99131748 |
| Marius Hobbhahnslice | Epoch (part-time research fellow)→ Epoch AI | → | Apollo Research (co-founder)→ Apollo Research | 2023 | Hobbhahn CV | af0df61432 |
| Adam Gleaveslice | Google DeepMind→ GOOGLE DEEPMIND | → | FAR.AI (co-founder); METR board→ FAR.AI · → METR | 2022 onward | Adam Gleave; METR | 5a1583e750 |
| Geoffrey Irvingslice | OpenAI; Google DeepMind→ OPENAI · → GOOGLE DEEPMIND | → | UK AI Security Institute (Apr 2024); Resolution (Jun 2026)→ RESOLUTION · → UK AISI | 2024 to 2026 | UK AISI; naml.us | a1541c890d |
| Jade Leungslice | GovAI; OpenAI (Governance Lead)→ GOVAI (CENTRE for the GOVERNANCE of AI) · → OPENAI | → | UK AI Safety Institute (Oct 2023); Prime Minister's AI adviser; Vice-Chair→ UK AISI | 2023 to 2026 | GovAI; GOV.UK | 735ace6e3a |
| Jeffrey Ladishslice | Anthropic's information security program (to 2022)→ ANTHROPIC | → | Palisade Research (Executive Director)→ Palisade Research | 2022 to 2023 | Palisade; Foresight Institute | 35ab51dfa9 |
| Jacob Steinhardtslice | OpenAI and Open Philanthropy (post-doc year)→ OPENAI · → OPEN PHILANTHROPY (COEFFICIENT GIVING) | → | UC Berkeley; Transluce (co-founder, Oct 2024)→ Transluce | 2024 | Berkeley; Transluce | 5030e97ba6 |
| Conrad Stoszslice | US Center for AI Standards and Innovation (led it)→ US CAISI | → | Transluce (Head of Governance)→ Transluce | 22 Oct 2025 | Transluce | 04210710e1 |
| J.J. Allaireslice | UK AI Security Institute (technical staff, Nov 2024)→ UK AISI | → | Meridian Labs (co-founder); visiting researcher at US CAISI→ MERIDIAN LABS · → US CAISI | 2025 | UK AISI; Meridian Labs | d61b436c68 |
| Alexandra Abbasslice | Arcadia Impact's AI Safety Engineering Taskforce (co-founder, 2024)no profile | → | Meridian Labs→ MERIDIAN LABS | 2025 to 2026 | Manifund; Meridian Labs | a107428208 |
| Kai Fronsdalslice | Anthropic Fellow; MATS→ ANTHROPIC · → MATS RESEARCH | → | Meridian Labs→ MERIDIAN LABS | 2025 to 2026 | Meridian Labs | d3bb4f226c |
| Kyle Scottslice | METR (Operations Manager)→ METR | → | Palisade Research board; ARC Treasurer→ Palisade Research · → ARC | by 2026 | Palisade; ARC | 7a4af92c25 |
| Lawrence Chanslice | Redwood Research→ Redwood Research | → | FAR.AI board→ FAR.AI | by 2026 | FAR.AI | 7d5f3ff2d3 |
| Bri Treeceslice | Center for AI Safety (COO)→ CENTER for AI SAFETY | → | FAR.AI board→ FAR.AI | by 2026 | FAR.AI | db5988702e |
| Ben Goldhaberslice | FAR.AI (Director of FAR.Labs)→ FAR.AI | → | Coefficient Giving→ OPEN PHILANTHROPY (COEFFICIENT GIVING) | Apr 2026 | Coefficient Giving | 6b1a9549ae |
| Erin Robertsonslice | Arcadia Impact (founder, co-director)no profile | → | Coefficient Giving→ OPEN PHILANTHROPY (COEFFICIENT GIVING) | by 2026 | Coefficient Giving | d9d5220e22 |
| Holden Karnofskyslice | Open Philanthropy (co-CEO, to 29 Apr 2024); ARC board (to Apr 2023)→ OPEN PHILANTHROPY (COEFFICIENT GIVING) · → ARC | → | Anthropic (technical staff, Jan 2025, reported)→ ANTHROPIC | 2024 to 2025 | EA Forum; Fortune | 70aff7d6b7 |
| Luke Muehlhauserslice | Open Philanthropy staff since 2015→ OPEN PHILANTHROPY (COEFFICIENT GIVING) | → | Anthropic board (to 29 May 2024); Coefficient Giving managing director→ ANTHROPIC · → OPEN PHILANTHROPY (COEFFICIENT GIVING) | 2024 | Anthropic; Coefficient Giving | 17709ea976 |
| Helen Tonerslice | Open Philanthropy (senior research analyst)→ OPEN PHILANTHROPY (COEFFICIENT GIVING) | → | Georgetown CSET (2019); OpenAI board (Sep 2021 to Nov 2023)→ OPENAI | 2019 to 2023 | CSET; OpenAI | 199039e038 |
| Alexandr Wangslice | Scale AI (CEO; remains Chairman)→ Scale AI | → | Meta (Chief AI Officer)→ META | Jun 2025 | Meta; Scale AI | 629f007abb |
From profiled 30 of 32To profiled 31 of 32Person profiled 27 of 32Dated 32 of 32
The instruments building the market as of 2026-10-17
The statutes, orders, tenders and standards that create demand for third-party evaluation, each with the status its record states and the dated milestones the pages carry. The next milestone is marked against today's date; a page that has nothing left ahead says so. An instrument names no evaluator, lab or person, so only the search box and the date window reach this board.
- California · statute chaptered
California SB 813 (Chapter 179, Statutes of 2026)
- AVERI's own legislative survey of 20 April 2026 lists SB 813 as "Create a voluntary Independent Verificatio
- California SB 813 (Chapter 179, Statutes of 2026), sponsored by Fathom and chaptered 9 September 2026, has the Government Operations Agency designate independent
- California: AVERI advisor Gillian Hadfield was named on 23 September 2026 to the Governor's group of experts advising on recommendati
- nstitution, nonprofit or corporation) the agency finds to have demonstrated expertise, by 1 January 2028.
- California · statute chaptered
California AB 1405 (Chapter 178)
- the companion AB 1405 (Chapter 178) creates an AI Auditor Registry with annual fees from 1 January 2029.
- California · executive order issued
California Executive Order N-9-26
- Executive Order N-9-26 of 18 September 2026 moves the criteria deadline to 1 May 2027.
- Executive Order N-9-26 of 18 September 2026 moves the criteria deadline to 1 May 2027.
- California · statute chaptered
California SB 53
- The Assembly Privacy committee's analysis of SB 53 (as amended 8 July 2025) says the bill "seeks to implement the recommendations of t
- Illinois · statute enacted
Illinois SB 315 (Public Act 104-0538)
- ILGA lists SB 315 as filed in January 2025 as a technical amendment to the Predatory Loan Prevention A
- 5/SB 3261) that it describes as the predecessor of SB 315, which the Governor approved on 6 July 2026, mandating annual independent third-party audits.
- ernor Approved" 6 July 2026, "Effective Date January 1, 2027" . - The pilot. "In July and August 2026, AVERI conducted the world’s first double-blind evaluation
- What the enrolled Illinois text requires. Public Act 104-0538 says "This Act takes effect January 1, 2027" but starts the audit later: "Beginning on January 1, 2028
- 38 says "This Act takes effect January 1, 2027" but starts the audit later: "Beginning on January 1, 2028 or 90 days after a developer first qualifies as a large fro
Illinois General Assembly · Public Act 104-0538, as enrolled · ILGA, HB 4705 · ILGA, SB 3261 b5369198d4
- Illinois · bill endorsed
Illinois HB 4705 / SB 3261
- IRS exemption issued July 2025 . - 15 January 2026. Launches with the flagship paper, "coa
- IRS exemption issued July 2025 . - 15 January 2026. Launches with the flagship paper, "coauthored with dozens
- Didech's (HB 4705, re-referred to Rules under Rule 19(a) on 27 March 2026) and Sen.
- nt to the Predatory Loan Prevention Act and rewritten by Senate floor amendment on 11 and 19 May 2026, and lists HB 4705 and SB 3261, the bills AVERI endorsed, a
- Edly-Allen's (SB 3261, re-referred to Assignments under Rule 3-9(a) on 22 May 2026).
Illinois General Assembly · ILGA, HB 4705 · ILGA, SB 3261 040a245db3
- United States · bill introduced
H.R. 9925
- ional Transparency, Independent Evaluation, and Reporting Act, introduced in the House on 23 July 2026.
- European Union · tender signed
EU AI Office tender EC-CNECT/2025/OP/0032
- g. Its technical-assistance tender EC-CNECT/2025/OP/0032 awarded six lots in November and December 2025, EUR 7,373,017.50 in all:.
- a group of EquiStamp, METR and Epoch AI for EUR 1,167,484 over 36 months, contract signed 15 December 2025.
- AI Evaluator Forum · standard published
AEF-1, Minimum Operating Conditions for Independent Third Party AI Evaluations
- e Intelligence Project and Princeton's Holistic Agent Leaderboard, announced the Forum in December 2025 with AEF-1, "Minimum Operating Conditions for Independent T
- -1, "Minimum Operating Conditions for Independent Third Party AI Evaluations" (Version 1, 4 December 2025), chaired by Conrad Stosz of Transluce with Patricia Paskov
AEF-1 · AI Evaluator Forum, members · AEF-1 · AI Evaluator Forum, letter fea91b8fa3
Dated milestones 9 of 9Status read 9 of 9
Filings on record as of 2026-10-17
For each evaluator whose profile carries an EIN: the registry, the latest return the profile has read, and the date it was read. The filings watcher asks the registry weekly whether a newer return exists; what it finds waits in a review queue until someone has read the return.
| Organization | Registry | Id | Latest return read | Read on | Record | Id |
|---|---|---|---|---|---|---|
| METRdossier · slice | IRS (via ProPublica Nonprofit Explorer) | 99-1219864 | no return read yet | 2026-10-17 | ProPublica Nonprofit Explorer | 9a5ad5e313 |
| SecureBiodossier · slice | IRS (via ProPublica Nonprofit Explorer) | 88-3068255 | FY2024 Form 990 | 2026-10-17 | ProPublica Nonprofit Explorer | 92db144676 |
| Redwood Researchdossier · slice | IRS (via ProPublica Nonprofit Explorer) | 87-1702255 | the 2024 return | 2026-10-17 | ProPublica Nonprofit Explorer | 6906668a4e |
| Epoch AIdossier · slice | IRS (via ProPublica Nonprofit Explorer) | 99-4050541 | FY2024 Form 990-EZ | 2026-10-17 | ProPublica Nonprofit Explorer | 07d1145c35 |
| Palisade Researchdossier · slice | IRS (via ProPublica Nonprofit Explorer) | 93-1591014 | for fiscal 2024 | 2026-10-17 | ProPublica Nonprofit Explorer | 46e1411939 |
| Translucedossier · slice | IRS (via ProPublica Nonprofit Explorer) | 99-3262331 | fiscal year ending December 2024 | 2026-10-14 | ProPublica Nonprofit Explorer | 846040f3b5 |
| AVERIdossier · slice | IRS (via ProPublica Nonprofit Explorer) | 38-4354994 | no Form 990 data for Ai Verification And Evalu | 2026-10-16 | ProPublica Nonprofit Explorer | e69a9a1e66 |
| Humane Intelligencedossier · slice | IRS (via ProPublica Nonprofit Explorer) | 93-1491311 | Form 990 for the year ending December 2024 | 2026-10-16 | ProPublica Nonprofit Explorer | aa187e0775 |
| Collective Intelligence Project (CIP)dossier · slice | IRS (via ProPublica Nonprofit Explorer) | 92-0327339 | the 2025 return | 2026-10-16 | ProPublica Nonprofit Explorer | 64bb527144 |
| FAR.AIdossier · slice | IRS (via ProPublica Nonprofit Explorer) | 92-0692207 | Form 990 for 2024 | 2026-10-14 | ProPublica Nonprofit Explorer | 30d7f460d2 |
| Apart Researchdossier · slice | IRS (via ProPublica Nonprofit Explorer) | 41-5214746 | no Form 990 data, so the filing record that ot | 2026-10-16 | ProPublica Nonprofit Explorer | 3a56b90e72 |
| ALIGNMENT RESEARCH CENTER (ARC)dossier · slice | IRS (via ProPublica Nonprofit Explorer) | 86-3605182 | FY2024 | 2026-10-14 | ProPublica Nonprofit Explorer | c1424a216d |
| MERIDIAN LABSdossier · slice | IRS (via ProPublica Nonprofit Explorer) | 33-3145457 | no Form 990 data | 2026-10-16 | ProPublica Nonprofit Explorer | bb6d60e73a |
Latest return read 12 of 13
What happened afterwards 18 rows, as of 2026-10-17
One row is one event a party published after an engagement: an incident on a model an evaluator had tested, a correction or added disclosure, a later finding that cut against an earlier one, a lab's statement that it acted or did not act on the finding, or a staff move from an evaluator to a lab it graded (read off the moves table). The evaluators a row names are the ones its source names; an incident also reaches every evaluator whose engagement covered the model. Each evaluator's own view, with the terms of each engagement, is its evaluator record.
| Kind | Evaluators | Developer | Model | What the record states | When | Source | Id |
|---|---|---|---|---|---|---|---|
| override | Apollo Research | OpenAI | o1 | "Apollo found that o1 has the capability to do basic in-context scheming (i.e. where the goal and knowledge of misalignment are acquired in context) and used scheming as a strategy in the scenarios within Apollo Research's evaluation suite"; "Apollo Research recommends monitoring for in-chain-of-thought scheming during deployment in agentic high-stakes settings such as automated AI research for next-generation frontier models"; the lab: "Carrying out monitoring for such purposes is an ongoing area of research and has various open challenges" | 21 December 2024 | OpenAI, o1 system card | 177dcc7fc6 |
| staff move | ARC | Anthropic | no model named | Holden Karnofsky: ARC board (to Apr 2023) -> Anthropic (technical staff, Jan 2025, reported)Holden Karnofsky | Jan 2025 | EA Forum; Fortune | fca6706722 |
| acted | Apollo Research | Anthropic | Claude Opus 4 | Apollo's summary in the card: "schemes and deceives at such high rates that we advise against deploying this model either internally or externally"; the lab: "These findings largely mirrored observations we had made internally about this early snapshot"; "This issue had not yet been mitigated as of the snapshot that they tested" | May 2025 | Anthropic, Claude Opus 4 and Sonnet 4 system card | 901a9b4ef1 |
| staff move | Scale AI | Meta | no model named | Alexandr Wang: Scale AI (CEO; remains Chairman) -> Meta (Chief AI Officer)Alexandr Wang | Jun 2025 | Meta; Scale AI | b3823f9ffa |
| acted | US CAISI | OpenAI | no model named | The lab's account of the institute's product red-teaming: CAISI "identified two novel security vulnerabilities in ChatGPT Agent"; "These attacks were immediately reported to OpenAI, and fixed by OpenAI within one business day" | 12 September 2025 | OpenAI, CAISI and UK AISI update | 8fe6218c20 |
| acted | Andon Labs | Anthropic | Opus 4.8 | In the section on Andon's testing: "Claude Opus 4.7, for example, had training that focused on business skills and robustness against adversarial agents, but we discovered that this training inadvertently contributed to misaligned behavior including dishonesty. We therefore removed it for Opus 4.8" | 28 May 2026 | Anthropic, Claude Opus 4.8 system card | 9157edf5e7 |
| correction | Epoch AI | OpenAI | FrontierMath | "On 2026-06-12, we released a major update, addressing errors in 42% of problems"; "We released v2, which corrected 123 problems in Tiers 1-3 and 12 problems in Tier 4. We also removed 5 problems from Tiers 1-3 and 7 problems from Tier 4" | 12 June 2026 | Epoch AI, FrontierMath Tier 4 (v2) | f2ba88ddba |
| incident | Irregularvia the model: Andon Labs, Deloitte, Frontier Design, Gray Swan, METR, UK AISI | Anthropic | Mythos 5; Opus 4.7 | "we identified three incidents in which a model accessed the internet from within or while interacting with the evaluation environment of Irregular, one of our third-party evaluation partners, and then gained unauthorized access to the production infrastructure of three different organizations"; "The incidents involved three different Claude models: Opus 4.7, Mythos 5, and an internal research test model"; "We notified our evaluation partner Irregular and the three affected organizations on Monday, July 27" | 30 July 2026 | Anthropic, investigating three incidents | dd4457ea07 |
| contradiction | UK AISI | Anthropic | Mythos 5 | Before release, in the system card: "[Claude Mythos 5] does not compromise safety research in our unprompted evaluations" ; after its own incident: "the activity undertaken by the agent show signs of novel, potentially deceptive behaviours, and were to an extent and severity we did not anticipate" | 4 August 2026 | AISI, incident report; Anthropic, Claude Fable 5 and Mythos 5 system card | c69cbb940f |
| incident | UK AISI; METRvia the model: Andon Labs, Deloitte, Frontier Design, Gray Swan, METR, UK AISI | Anthropic | Mythos 5 | The Institute's own cyber-testing incident: "in 10 of the 122 runs, we identified 19 cases where an agent had taken distinct actions beyond the scope of the testing parameters. 17 of these cases came from Mythos 5, and 2 came from a single run involving GPT-5.6 Sol"; "The developers' cyber classifiers were deliberately switched off"; "This behaviour occurred between July 25th and July 28th 2026"; "We also intend to work with METR (Model Evaluation and Threat Research) to conduct an independent third-party review" | 4 August 2026 | AISI, incident report | 1a39056dce |
| incident | UK AISIvia the model: Apollo Research, Irregular, METR, SecureBio, UK AISI | OpenAI | GPT-5.6 Sol | In the same runs, "2 actions involving OpenAI's GPT-5.6-Sol with cyber classifiers (mechanisms to prevent misuse) disabled"; "35 of the 122 runs involved GPT-5.6 Sol" | 4 August 2026 | AISI, incident report | f1b5d7b991 |
| incident | METR; Redwood Researchvia the model: Apollo Research, Irregular, METR, SecureBio, UK AISI | OpenAI | GPT-5.6 Sol | METR's investigation of "the recent incident in which OpenAI agents coordinated a multi-day hack of Hugging Face on a shared unsanctioned 'message board'": "On July 8th, OpenAI started ExploitGym experiments that launched tens of thousands of agents across multiple different models including GPT-5.6 Sol and a highly-persistent internal model"; the investigators "worked on premises at OpenAI over a total of six days"; "Per our standard policy, we did not take payment from OpenAI for this independent assessment"; "OpenAI was able to redact any non-public information from this post" | 26 August 2026 | METR, Hugging Face incident investigation | fabf586e95 |
| staff move | Epoch AI; MECHANIZE | Google DeepMind | no model named | Tamay Besiroglu: Epoch AI (co-founder, associate director); Mechanize (co-founder, Apr 2025) -> Google DeepMind (research scientist, by Sep 2026)Tamay Besiroglu | Sep 2026 | Epoch AI; Mechanize; his post | 75f848d85a |
| correction | Irregular | Anthropic | Mythos 5; Opus 4.7 | The lab's revised count and reading of the July incidents: "This missed a set of transcripts that also turned out to have internet access; we identified these in August while assembling transcripts to share with METR"; "All four incidents occurred during cybersecurity evaluations built by the same evaluation partner"; "we believe Claude's behavior reflects two forms of misalignment: Biased reasoning, in which models selectively interpret evidence in ways that favor justifying their actions; Recklessness, in which models have a propensity to keep trying to solve their task, even when this could lead to harm" | 9 September 2026 | Anthropic, alignment assessment of recent cybersecurity incidents | 13fbc56ae6 |
| staff move | ARC; US CAISI | OpenAI | no model named | Paul Christiano: ARC; US AI Safety Institute (Apr 2024); CAISI senior technical adviser; back to ARC as executive director (Aug 2026) -> OpenAI Foundation board (9 Sep 2026)Paul Christiano | 9 Sep 2026 | NIST; OpenAI Foundation; ARC | 47b741082c |
| correction | METR | OpenAI | no model named | Disclosures added to the published investigation: "This post was edited on 9/13/2026 to add footnotes on Ajeya Cotra and Ryan Greenblatt. On Sep 9th (after this report was completed and published on Aug 26th), Ajeya Cotra's spouse Paul Christiano joined OpenAI's Safety and Security Committee"; "Ryan Greenblatt is the domestic partner of Beth Barnes, METR's CEO. Beth Barnes was not involved in the decision to engage Ryan Greenblatt, nor was she directly involved in this investigation" | 13 September 2026 | METR, Hugging Face incident investigation | 1b199404af |
| correction | Epoch AI | OpenAI | GPT-6 Astra | Its chief executive on the model's Epoch Capabilities Index score: "An update on Astra's ECI score. This is a major step down from the original score we reported (but still within the uncertainty range we expected). It is explained by lower performance on long-horizon software engineering tasks" | 16 September 2026 | Jaime Sevilla on X | 5d33e90812 |
| override | Faculty | Google DeepMind | Gemini 4 Argon | "The external evaluator found that the model offered a substantial capability uplift for certain threat actors across the domains"; the lab: "Overall, these findings did not change our risk determination when reviewed alongside internal evaluation results"; on cyber, "These results were reviewed by internal SMEs and did not change our Cyber CCL determination, because they were consistent with our understanding of model capabilities" | October 2026 | Google DeepMind, Gemini 4 Argon safety report | 880899df06 |
Dated 18 of 18Model named 12 of 18Tied to an engagement 13 of 18Evaluator named by the record 18 of 18
The evaluators 43 organizations, as of 2026-10-23
15 nonprofit · 21 for-profit · 3 government · 2 academic · 2 unstated. Type is read from the roster's own money column and the profile's status line. Founded and HQ are what the profile states, with the clause they were read from; a registry id appears only where a cited page prints it, and a for-profit without one is shown as such. Each evaluator's record page gathers its engagements, terms, findings and what followed.
Founded 27 of 43HQ 34 of 43Registry id 15 of 43Graph node 39 of 43