LatchBio
What it was engaged to test, on what terms, what it found in the document's own words, and what happened afterwards. Every line is a record; every record links to the document it was read from.
- Founded2021
- HQSan Francisco
- Registry idnone cited (company)
- TestsBiology benchmarks and biosecurity audits: SpatialBench Verified and SingleCellBench in six Anthropic cards (the Haiku 5.5 card cross-references them for its CB-1 threat model in place of BioMysteryBench); BioSecBench analysis of Grok 4.6 published by xAI on 1 September 2026 and its benchmark cited in the Grok 4.7 post and card.
- Money, as the roster puts itVenture-backed, San Francisco, founded 2021: $5M seed led by Lux Capital (October 2021); "$28M Series A from Lux Capital, Coatue, and General Catalyst"; also Fifty Years, Haystack, Jeff Dean, Alec Nielsen and George Church; no lab payment stated in any document.
- Filesdossier · observatory, sliced to it · graph
- Span of engagements
- 9 June 2026 – 7 October 2026
- Terms and findings
- 4 of 9 documents state a term · 4 quote a finding
- Money in, US dollars, stated single-recipient rows
- no stated figure 0 in · 0 out
- Seats, by side
- no seat names it
- Afterwards
- no later event on record
Each count is the number of store records that name this evaluator. An engagement is one lab document naming it for one model; an outcome is an event on the record afterwards; the money, seats and moves are the observatory's rows filtered to it. A cell that cannot be filled says not stated, and each table ends with how many of its rows state each field.
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.
What it was engaged to test 9 engagements, Anthropic, xAI
In date order, each from the lab document that names the evaluator. Under each: the four terms that document states (access, model stage, publication, payment) and the finding in its words. A term the document is silent on reads not stated; the tags beside the terms are read off their words and are what the console filters on.
- Document: Claude Fable 5 and Mythos 5 system card
- Access windownot stated
- Model stagenot stated
- Publicationnot stated
- Paymentnot stated
Finding: not stated in the document
- Document: Claude Opus 5 system card
- Access windownot stated
- Model stagenot stated
- Publicationnot stated
- Paymentnot stated
Finding: not stated in the document
Sources: Claude Opus 5 system card b916b9369d - Access window"Claude has access to a bash tool for code execution, a file editor, and a pre-specified set of installed packages"
- Model stagenot stated
- Publication"Although many of these evaluations are not publicly released, we briefly describe each one below"; "We did not run private benchmarks (such as LatchBio's) on GPT or Gemini"
- Paymentnot stated
Finding, in the document's words"On SpatialBench Verified, Claude Mythos 5.1 achieved the top score at 77.6%, ahead of Claude Opus 5 at 72.5%"
- Document: xAI, Grok 4.6 biosecurity post
- Access window"LatchBio performed testing of the latest-available version of the model on our full suite of biological-capability tests"
- Model stage"The currently-served version of Grok 4.6", tested "Following the initial release of Grok 4.6"
- Publication"Today, LatchBio published an independent analysis of Grok 4.6"; "LatchBio's full methods and per-model scores are presented at benchmarks.bio"
- Paymentnot stated
Finding, in the document's words"On LatchBio's BioSecBench-Refusal suite, Grok 4.6 was the strongest model tested at refusing disguised and hazardous tasks while still completing routine biological work. It was the only system to score above 50% on both measures"
Sources: xAI, Grok 4.6 biosecurity post; LatchBio; LatchBio 876d37d51e - Document: xAI, Grok 4.7 model card
- Access window"GPT-6 Astra and Opus 5 are LatchBio runs at max effort"
- Model stage"Grok 4.7 and Grok 4.6 are run with the Grok Build harness"
- Publication"The LatchBio capability benchmark suite is a live suite, so the model card should freeze the evaluated snapshot and coverage"
- Paymentnot stated
Finding, in the document's wordsRefusal: "Grok 4.6 (high)" 45.6%, "Grok 4.7 (xhigh)" 62.4%, "GPT-6 Astra (max)" 25.5%, "Opus 5 (max)" 31.8%; "Provider-side filters can dominate refusal behavior, and the restricted task sets characterize the tested checkpoint and harness rather than establishing general biological safety or reliability"
Sources: xAI, Grok 4.7 model card 2d870b332a - Document: Introducing Grok 4.7
- Access windownot stated
- Model stagenot stated
- Publicationnot stated
- Paymentnot stated
Finding: not stated in the document
Sources: Introducing Grok 4.7 d9c2203b2b - Document: Claude Opus 5.5 system card
- Access windownot stated
- Model stagenot stated
- Publicationnot stated
- Paymentnot stated
Finding: not stated in the document
Sources: Claude Opus 5.5 system card d49850fc00 - Document: Claude Sonnet 5.5 system card
- Access windownot stated
- Model stagenot stated
- Publicationnot stated
- Paymentnot stated
Finding: not stated in the document
Sources: Claude Sonnet 5.5 system card 4bba9849a9 - Document: Claude Haiku 5.5 system card
- Access windownot stated
- Model stagenot stated
- Publicationnot stated
- Paymentnot stated
Finding, in the document's words"BioMysteryBench has been deprecated and removed from our panel of Life Sciences capabilities evaluations. In its place, we now cross-reference the LatchBio Bioinformatics suite to assess relevant capabilities in bioinformatics"; "Haiku 5.5 performed comparably to Sonnet 5, but remained weaker than Opus 5 on both LatchBio Bioinformatics evaluations"
Sources: Claude Haiku 5.5 system card 2be2d20015
Dated 9 of 9Model named 9 of 9Terms stated 4 of 9Finding quoted 4 of 9
The terms, as the documents state them
The same engagements as a grid, one row per document and model, so the silences line up: a document that gives an access window but no payment term, or a finding but no publication term, reads that way here.
| Document | Model | Access window | Model stage | Publication | Payment | Finding | Record |
|---|---|---|---|---|---|---|---|
| Claude Fable 5 and Mythos 5 system cardAnthropic · 9 June 2026 | Fable 5 / Mythos 5 | not stated | not stated | not stated | not stated | not stated | 2fe1588075 |
| Claude Opus 5 system cardAnthropic · 24 July 2026 | Opus 5 | not stated | not stated | not stated | not stated | not stated | b916b9369d |
| Claude Fable 5.1 and Mythos 5.1 system cardAnthropic · 1 September 2026 | Fable 5.1 | "Claude has access to a bash tool for code execution, a file editor, and a pre-specified set of installed packages" | not stated | "Although many of these evaluations are not publicly released, we briefly describe each one below"; "We did not run private benchmarks (such as LatchBio's) on GPT or Gemini" | not stated | quoted | ca03b34bdd |
| xAI, Grok 4.6 biosecurity postxAI · 1 September 2026 | Grok 4.6 | "LatchBio performed testing of the latest-available version of the model on our full suite of biological-capability tests" | "The currently-served version of Grok 4.6", tested "Following the initial release of Grok 4.6" | "Today, LatchBio published an independent analysis of Grok 4.6"; "LatchBio's full methods and per-model scores are presented at benchmarks.bio" | not stated | quoted | 876d37d51e |
| xAI, Grok 4.7 model cardxAI · 21 September 2026 | Grok 4.7 | "GPT-6 Astra and Opus 5 are LatchBio runs at max effort" | "Grok 4.7 and Grok 4.6 are run with the Grok Build harness" | "The LatchBio capability benchmark suite is a live suite, so the model card should freeze the evaluated snapshot and coverage" | not stated | quoted | 2d870b332a |
| Introducing Grok 4.7xAI · 21 September 2026 | Grok 4.7 | not stated | not stated | not stated | not stated | not stated | d9c2203b2b |
| Claude Opus 5.5 system cardAnthropic · 22 September 2026 | Opus 5.5 | not stated | not stated | not stated | not stated | not stated | d49850fc00 |
| Claude Sonnet 5.5 system cardAnthropic · 28 September 2026 | Sonnet 5.5 | not stated | not stated | not stated | not stated | not stated | 4bba9849a9 |
| Claude Haiku 5.5 system cardAnthropic · 7 October 2026 | Haiku 5.5 | not stated | not stated | not stated | not stated | quoted | 2be2d20015 |
Access window 3 of 9Model stage 2 of 9Publication 3 of 9Payment 0 of 9Finding, quoted 4 of 9
What happened afterwards 0 on record
Events a party published after an engagement: an incident on a model this evaluator's engagement covered, 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 move from this evaluator to a lab it graded. A row the source attaches to this evaluator by name is marked named; one that reaches it through the model it tested is marked via the model. The kind is what the record states, not why.
| Kind | Developer | Model | What the record states | When | Source | Id |
|---|---|---|---|---|---|---|
| no later event on record | ||||||
Dated 0 of 0Model named 0 of 0Tied to an engagement 0 of 0Evaluator named by the record 0 of 0
The money on record 0 rows
Rows of the observatory's money table that name this evaluator as recipient (in) or as funder (out), each as the record states it: a grant table line, a contract notice, a Form 990 schedule or the party's own disclosure. A figure quoted in prose is marked and never summed.
| Direction | Funder | Recipient | Amount, as recorded | When | Record | Source | Id |
|---|---|---|---|---|---|---|---|
| no money row names this evaluator | |||||||
Amount stated 0 of 0Dated 0 of 0Purpose stated 0 of 0Funder profiled 0 of 0Recipient on the roster 0 of 0
Seats and moves 0 seats · 0 moves
Seats read off a team page, a governance page, an appointment notice or a filing that name this organization, and the dated moves into or out of it, from the observatory's own tables.
| Person | Side | Seat, as the record states it | Role | Dates | Source | Id |
|---|---|---|---|---|---|---|
| no seat on record names this evaluator | ||||||
Role read 0 of 0Organization read 0 of 0Organization profiled 0 of 0Dated 0 of 0
Moves
| Person | From | To | When | Source | Id | |
|---|---|---|---|---|---|---|
| no move on record names this evaluator | ||||||
From profiled 0 of 0To profiled 0 of 0Person profiled 0 of 0Dated 0 of 0