Benchmark Almanac

LatchBio Capabilities v1.0

Claude Opus 5.5 58% leads since Sep 2026 Only in system cards

LatchBio cross-benchmark suite of agentic biological data analysis (transcriptomics, epigenomics, variants, pharmacology) with deterministic graders.

The V1 Capabilities score on benchmarks.bio is a weighted mean of 13 LatchBio benchmarks. They are the SpatialBench, scBench, EpiBench, VariantBench, MetagenomicsBench, TxBench, BioSecBench and OncologyBench sets. Legacy V0 scores are not comparable with V1. Fireworks lists the suite as BioBench.

Made by LatchBio Released 2026 Lab and data work Last in a system card Sep 2026 Last model added Oct 2026 Scores refreshed 2 Nov 2026
Level
Expert Specialists with years of training, often over hours of work
Scored
Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
Answers
Short answer A number, word or short string matched to a key.
Leaderboard benchmarks.bio
27
Models scored on its own leaderboard
1
System card reports it, from 1 lab

Claude Opus 5.5 leads with 58%

Best score per model from the benchmark's own leaderboard, read on 2 Nov 2026. Where a model ran with several settings or harnesses, we keep its best.

27 models. Scroll sideways for the rest.

  1. 58.0 Claude Opus 5.5
  2. 54.2 Claude Sonnet 5.5
  3. 53.0 Claude Fable 5.1
  4. 52.4 Gemini 4 Argon
  5. 51.3 GPT-6 Astra
  6. 50.6 GPT-6.1 Sol
  7. 50.0 Claude Opus 5
  8. 47.4 Grok 4.6
  9. 47.4 Grok 4.7
  10. 46.7 GPT-6 Sol
  11. 45.3 Claude Opus 4.8
  12. 43.4 GPT-6.1 Terra
  13. 43.3 GPT-5.6 Sol
  14. 42.1 DeepSeek V4.1 Flash
  15. 41.9 GPT-5.5
  16. 40.9 Claude Sonnet 5
  17. 40.8 Gemini 3.7 Flash
  18. 38.6 Claude Opus 4.7
  19. 37.7 Gemini 3.8 Flash
  20. 37.2 GPT-6 Luna
  21. 36.4 Kimi K3
  22. 34.9 Mistral Large 4
  23. 33.6 GLM-5.3
  24. 33.5 GPT-5.6 Luna
  25. 31.5 Claude Sonnet 4.6
  26. 24.2 Nemotron 3 Ultra
  27. 14.4 Nemotron 3.5 Lightning 30B-A3B
  1. 1 Claude Opus 5.5 LSVP 58%
  2. 2 Claude Sonnet 5.5 54.2%
  3. 3 Claude Fable 5.1 LSVP 53%
  4. 4 Gemini 4 Argon 52.4%
  5. 5 GPT-6 Astra 51.3%
  6. 6 GPT-6.1 Sol 50.6%
  7. 7 Claude Opus 5 50%
  8. 8 Grok 4.6 47.4%
  9. 9 Grok 4.7 47.4%
  10. 10 GPT-6 Sol 46.7%
  11. 11 Claude Opus 4.8 45.3%
  12. 12 GPT-6.1 Terra 43.4%
  13. 13 GPT-5.6 Sol 43.3%
  14. 14 DeepSeek V4.1 Flash 42.1%
  15. 15 GPT-5.5 41.9%
  16. 16 Claude Sonnet 5 40.9%
  17. 17 Gemini 3.7 Flash 40.8%
  18. 18 Claude Opus 4.7 38.6%
  19. 19 Gemini 3.8 Flash 37.7%
  20. 20 GPT-6 Luna 37.2%
  21. 21 Kimi K3 36.4%
  22. 22 Mistral Large 4 34.9%
  23. 23 GLM-5.3 33.6%
  24. 24 GPT-5.6 Luna 33.5%
  25. 25 Claude Sonnet 4.6 31.5%
  26. 26 Nemotron 3 Ultra 24.2%
  27. 27 Nemotron 3.5 Lightning 30B-A3B 14.4%
#ModelScoreSettingDateSource
1 Claude Opus 5.5 58% LSVP, Pi Sep 2026 Official leaderboard
2 Claude Sonnet 5.5 54.2% Pi Sep 2026 Official leaderboard
3 Claude Fable 5.1 53% LSVP, Pi Sep 2026 Official leaderboard
4 Gemini 4 Argon 52.4% Pi Sep 2026 Official leaderboard
5 GPT-6 Astra 51.3% Pi Sep 2026 Official leaderboard
6 GPT-6.1 Sol 50.6% Pi Sep 2026 Official leaderboard
7 Claude Opus 5 50% Pi Jul 2026 Official leaderboard
8 Grok 4.6 47.4% Pi Aug 2026 Official leaderboard
9 Grok 4.7 47.4% Pi Sep 2026 Official leaderboard
10 GPT-6 Sol 46.7% Pi Sep 2026 Official leaderboard
11 Claude Opus 4.8 45.3% Pi May 2026 Official leaderboard
12 GPT-6.1 Terra 43.4% Pi Oct 2026 Official leaderboard
13 GPT-5.6 Sol 43.3% Pi Jul 2026 Official leaderboard
14 DeepSeek V4.1 Flash 42.1% Pi Sep 2026 Official leaderboard
15 GPT-5.5 41.9% Pi Apr 2026 Official leaderboard
16 Claude Sonnet 5 40.9% Pi Jun 2026 Official leaderboard
17 Gemini 3.7 Flash 40.8% Pi Aug 2026 Official leaderboard
18 Claude Opus 4.7 38.6% Pi Apr 2026 Official leaderboard
19 Gemini 3.8 Flash 37.7% Pi Sep 2026 Official leaderboard
20 GPT-6 Luna 37.2% Pi Sep 2026 Official leaderboard
21 Kimi K3 36.4% Pi Jul 2026 Official leaderboard
22 Mistral Large 4 34.9% Pi Oct 2026 Official leaderboard
23 GLM-5.3 33.6% Pi Aug 2026 Official leaderboard
24 GPT-5.6 Luna 33.5% Codex Jul 2026 Official leaderboard
25 Claude Sonnet 4.6 31.5% Pi Feb 2026 Official leaderboard
26 Nemotron 3 Ultra 24.2% Pi Jun 2026 Official leaderboard
27 Nemotron 3.5 Lightning 30B-A3B 14.4% Pi Aug 2026 Official leaderboard

Labs report scores for 1 model in their cards. We plot every source in the charts below.

7 models have held number one since its release in Jan 2026

Lab claims count from the day the card came out, independent runs from the model’s release.

Anthropic OpenAI
Own leaderboard 7 leaders
Claude Opus 5.5 58%

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) Claude Opus 5.5 Claude Fable 5.1 Claude Opus 4.8 Claude Opus 4.7 Claude Sonnet 4.6 0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) Claude Opus 5.5 Claude Opus 5

Labs report it only in system cards

We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark. Some system cards do, below.

One system card reports it

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

DateDocumentScoreAs the document writes itSection
21 Sep 2026 Grok 4.7 model card 44.5 LatchBio Capabilities v1.0, overall score (mean of 11 capability benchmarks), xhigh (5.2) Capabilities

More in Science and reasoning

All 88 in Science and reasoning
  • GPQA Diamond In use 198 graduate-level 'Google-proof' multiple-choice questions in biology, chemistry and physics. Gemini 4 Argon 96.4%
  • Terminal-Bench-Science In use 70 scientific research-workflow tasks across five science areas, run in Harbor containers. Hidden tests give the resolution rate. GPT-6 Astra 68.1%
  • FrontierScience In use Models solve physics, chemistry, and biology problems in an Olympiad track graded by short answer and a Research track graded by 10-point rubrics. Intern-S2-397B 87%
  • BioMysteryBench In use Bioinformatics problems on anonymized real datasets that require real analysis with tools. Rubric-graded, with human-solvable and human-difficult splits. Claude Opus 5 90.1%
  • HiPhO In use Models solve theory problems from 2024 and 2025 physics olympiad exams, some with diagrams. Official marking schemes grade answers and steps, mapped to medal levels. Gemini 3 Pro 78.3%
  • LABBench2 In use 11 biology research sub-tasks (literature, figures, tables, patents, trials and more) with tool access. Macro-averaged accuracy. GPT-5.2 60.9%