Benchmark Almanac

LatchBio Capabilities V0

Claude Sonnet 5.5 51.3% leads since September 2026
Not in model cards

Agents complete biological data analysis tasks. The historical score averages the available benchmark scores equally and excludes missing scores.

LatchBio reports 46 model and harness combinations for V0. V1 changes both the evaluation sets and the weighting.

Made by
LatchBio
Kind
Genomics and life sciences
Level
Expert
Scored
Checked automatically
Answers
Agent task
Tasks
Public, plus a private test set
Measures
Capability
Language
English
In and out
Text, Tables → Text, Actions
Interaction
Agentic
Environment
Code sandbox
Items from
Another benchmark
Run by
A host
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
October 2026
Scores refreshed
November 21st, 2026
46
Entries on LatchBio’s leaderboard, 17 models we track

Claude Sonnet 5.5 leads with 51.3% Claude Sonnet 5.5 leads with 51.3%

Best score per entry from LatchBio’s leaderboard, read on November 21st, 2026. Where a model ran on its own with several settings, we keep its best. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. Agents built around a model, such as Pi + GPT-6 Astra, are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best.

  1. 1 Claude Sonnet 5.5 51.3%
  2. 2 Pi + GPT-6 Astra 48.7%
  3. 3 GPT-6 Astra 48.5%
  4. 4 Claude Opus 5 48%
  5. 5 Pi + Claude Opus 5 48%
  6. 6 Pi + Claude Sonnet 5.5 48%
  7. 7 Grok 4.6 45.4%
  8. 8 Pi + Claude Opus 5.5 45%
  9. 9 Pi + GPT-5.6 Sol 43.4%
  10. 10 Pi + Gemini 3.8 Flash 43.4%
  11. 11 Pi + Claude Opus 4.8 42.7%
  12. 12 Pi + GPT-5.6 Terra 42.4%
  13. 13 Pi + Grok 4.6 42.3%
  14. 14 Claude Opus 5.5 41.9%
  15. 15 GPT-5.6 Sol 41.8%
  16. 16 Pi + Gemini 3.5 Flash 41.7%
  17. 17 Pi + DeepSeek V4.1 Flash 41.6%
  18. 18 GPT-5.6 Terra 41.5%
  19. 19 Pi + Grok 4.5 41.3%
  20. 20 Claude Opus 4.8 41.1%
  21. 21 Claude Opus 4.7 39.8%
  22. 22 GPT-5.5 38.7%
  23. 23 Pi + GPT-5.5 37.8%
  24. 24 Pi + Claude Sonnet 5 37.5%
  25. 25 Claude Sonnet 5 37%
  26. 26 Pi + Gemini 3.7 Flash 36.5%
  27. 27 Pi + Kimi K3 35.3%
  28. 28 Grok 4.7 33.7%
  29. 29 Pi + Grok 4.7 33.5%
  30. 30 Pi + Gemini 4 Argon 33.3%
  31. 31 Pi + GPT-5.6 Luna 31%
  32. 32 GPT-6 Sol 30.4%
  33. 33 Pi + GPT-6 Sol 30.3%
  34. 34 GPT-5.6 Luna 29.7%
  35. 35 Pi + Mistral Large 4 25.3%
  36. 36 Pi + Muse Wonder 24%
  37. 37 GPT-6 Luna 23.5%
  38. 38 Pi + Claude Sonnet 4.6 22.3%
  39. 39 Pi + GPT-6 Luna 21.4%
  40. 40 Pi + GLM-5.3 17.3%
  41. 41 Claude Sonnet 4.6 16.8%
  42. 42 Pi + GPT-6.1 Terra 16%
  43. 43 Pi + Muse Spark Open 12%
  44. 44 GPT-6.1 Terra 9.30%
  45. 45 Pi + Mercury 3 8.00%
  46. 46 Pi + Beam 1.0 1.30%

Source: LatchBio’s leaderboard

#EntryScoreSettingSubmitted byDateSource
1 Claude Sonnet 5.5 51.3% Claude Code Anthropic Sep 2026 Official leaderboard
2 Pi + GPT-6 Astra 48.7% OpenAI Sep 2026 Official leaderboard
3 GPT-6 Astra 48.5% Codex OpenAI Sep 2026 Official leaderboard
4 Claude Opus 5 48% Claude Code Anthropic Jul 2026 Official leaderboard
5 Pi + Claude Opus 5 48% Anthropic Jul 2026 Official leaderboard
6 Pi + Claude Sonnet 5.5 48% Anthropic Sep 2026 Official leaderboard
7 Grok 4.6 45.4% Grok Build SpaceXAI Aug 2026 Official leaderboard
8 Pi + Claude Opus 5.5 45% Anthropic Sep 2026 Official leaderboard
9 Pi + GPT-5.6 Sol 43.4% OpenAI Jul 2026 Official leaderboard
10 Pi + Gemini 3.8 Flash 43.4% Google Sep 2026 Official leaderboard
11 Pi + Claude Opus 4.8 42.7% Anthropic May 2026 Official leaderboard
12 Pi + GPT-5.6 Terra 42.4% OpenAI Jul 2026 Official leaderboard
13 Pi + Grok 4.6 42.3% SpaceXAI Aug 2026 Official leaderboard
14 Claude Opus 5.5 41.9% Claude Code Anthropic Sep 2026 Official leaderboard
15 GPT-5.6 Sol 41.8% Codex OpenAI Jul 2026 Official leaderboard
16 Pi + Gemini 3.5 Flash 41.7% Google May 2026 Official leaderboard
17 Pi + DeepSeek V4.1 Flash 41.6% DeepSeek Sep 2026 Official leaderboard
18 GPT-5.6 Terra 41.5% Codex OpenAI Jul 2026 Official leaderboard
19 Pi + Grok 4.5 41.3% SpaceXAI Jul 2026 Official leaderboard
20 Claude Opus 4.8 41.1% Claude Code Anthropic May 2026 Official leaderboard
21 Claude Opus 4.7 39.8% Claude Code Anthropic Apr 2026 Official leaderboard
22 GPT-5.5 38.7% Codex OpenAI Apr 2026 Official leaderboard
23 Pi + GPT-5.5 37.8% OpenAI Apr 2026 Official leaderboard
24 Pi + Claude Sonnet 5 37.5% Anthropic Jun 2026 Official leaderboard
25 Claude Sonnet 5 37% Claude Code Anthropic Jun 2026 Official leaderboard
26 Pi + Gemini 3.7 Flash 36.5% Google Aug 2026 Official leaderboard
27 Pi + Kimi K3 35.3% Moonshot Jul 2026 Official leaderboard
28 Grok 4.7 33.7% Grok Build SpaceXAI Sep 2026 Official leaderboard
29 Pi + Grok 4.7 33.5% SpaceXAI Sep 2026 Official leaderboard
30 Pi + Gemini 4 Argon 33.3% Google Sep 2026 Official leaderboard
31 Pi + GPT-5.6 Luna 31% OpenAI Jul 2026 Official leaderboard
32 GPT-6 Sol 30.4% Codex OpenAI Sep 2026 Official leaderboard
33 Pi + GPT-6 Sol 30.3% OpenAI Sep 2026 Official leaderboard
34 GPT-5.6 Luna 29.7% Codex OpenAI Jul 2026 Official leaderboard
35 Pi + Mistral Large 4 25.3% Mistral Oct 2026 Official leaderboard
36 Pi + Muse Wonder 24% Meta Oct 2026 Official leaderboard
37 GPT-6 Luna 23.5% Codex OpenAI Sep 2026 Official leaderboard
38 Pi + Claude Sonnet 4.6 22.3% Anthropic Feb 2026 Official leaderboard
39 Pi + GPT-6 Luna 21.4% OpenAI Sep 2026 Official leaderboard
40 Pi + GLM-5.3 17.3% Z.ai Aug 2026 Official leaderboard
41 Claude Sonnet 4.6 16.8% Claude Code Anthropic Feb 2026 Official leaderboard
42 Pi + GPT-6.1 Terra 16% OpenAI Oct 2026 Official leaderboard
43 Pi + Muse Spark Open 12% Meta Oct 2026 Official leaderboard
44 GPT-6.1 Terra 9.30% Codex OpenAI Oct 2026 Official leaderboard
45 Pi + Mercury 3 8.00% Inception Oct 2026 Official leaderboard
46 Pi + Beam 1.0 1.30% Reflection Oct 2026 Official leaderboard
#ModelScoreSettingSubmitted byDateSource
1 Claude Sonnet 5.5 51.3% Claude Code Anthropic Sep 2026 Official leaderboard
2 GPT-6 Astra 48.5% Codex OpenAI Sep 2026 Official leaderboard
3 Claude Opus 5 48% Claude Code Anthropic Jul 2026 Official leaderboard
4 Grok 4.6 45.4% Grok Build SpaceXAI Aug 2026 Official leaderboard
5 Claude Opus 5.5 41.9% Claude Code Anthropic Sep 2026 Official leaderboard
6 GPT-5.6 Sol 41.8% Codex OpenAI Jul 2026 Official leaderboard
7 GPT-5.6 Terra 41.5% Codex OpenAI Jul 2026 Official leaderboard
8 Claude Opus 4.8 41.1% Claude Code Anthropic May 2026 Official leaderboard
9 Claude Opus 4.7 39.8% Claude Code Anthropic Apr 2026 Official leaderboard
10 GPT-5.5 38.7% Codex OpenAI Apr 2026 Official leaderboard
11 Claude Sonnet 5 37% Claude Code Anthropic Jun 2026 Official leaderboard
12 Grok 4.7 33.7% Grok Build SpaceXAI Sep 2026 Official leaderboard
13 GPT-6 Sol 30.4% Codex OpenAI Sep 2026 Official leaderboard
14 GPT-5.6 Luna 29.7% Codex OpenAI Jul 2026 Official leaderboard
15 GPT-6 Luna 23.5% Codex OpenAI Sep 2026 Official leaderboard
16 Claude Sonnet 4.6 16.8% Claude Code Anthropic Feb 2026 Official leaderboard
17 GPT-6.1 Terra 9.30% Codex OpenAI Oct 2026 Official leaderboard

Rankings by part

We read 13 parts from LatchBio’s leaderboard, each ranking models on one part of the test, such as spatialbench, spatialbench-long and scbench. GPT-6 Astra leads 4 of them, and 5 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

spatialbench

17 models we track have a score here.

  1. 1 Grok 4.6 ↑3 77.4%
  2. 2 GPT-5.6 Sol ↑4 74.5%
  3. 3 Claude Opus 5 71%
  4. 4 GPT-6 Astra ↓2 70.1%
  5. 5 GPT-5.6 Terra ↑2 69.6%
  6. 6 Claude Opus 4.8 ↑2 69%
  7. 7 GPT-5.5 ↑3 64.9%
  8. 8 Claude Sonnet 5 ↑3 64.3%
  9. 9 Claude Opus 4.7 64.1%
  10. 10 GPT-5.6 Luna ↑4 60.6%

7 models have held number one on LatchBio’s leaderboard since 2023

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

Anthropic OpenAI
LatchBio leaderboard 7 leaders

Source: LatchBio’s leaderboard

Scores over time

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

Plot with other benchmarks
On LatchBio’s leaderboard Best so far
0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) Claude Sonnet 5.5 Claude Opus 5 GPT-5.6 Sol Claude Opus 4.7 Claude Sonnet 4.6 0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) Claude Sonnet 5.5 GPT-5.6 Sol

Source: LatchBio’s leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

Versions

More in Biology and chemistry

All 137 in Biology and chemistry
  • GeneBench Pro In use Agents analyze messy simulated genomics and biology datasets to estimate a decision-relevant quantity across 129 problems. Scored automatically against known answers. GPT-5.6 Sol 31.5%
  • MedChemBench (OpenAI) In use Models work through 99 medicinal chemistry problems drawn from recent papers and patents. Each task mixes prose with structures, assay tables and plots, then asks for the program’s next decision. GPT-6 Astra 83.8%
  • SUPERChem In use Models solve 500 expert-written chemistry reasoning problems, offered in multimodal and text-only forms. Expert solution paths also allow a reasoning path score. Tencent Hy4 preview 66.4%
  • LifeSciBench In use Models complete 750 expert-written life science research tasks, often with attached figures, tables, or sequence files. Graded against expert rubrics as score and pass rate. GPT-Rosalind-6 61.3%
  • ProtocolQA Open-Ended In use Models correct errors in published lab protocols using open-ended questions adapted from LAB-Bench ProtocolQA. Scored as pass@1. GPT-6 Cyber 47.2%
  • MolecularIQ In use Models answer questions about molecular graphs, such as counting, indexing and constrained generation, and a symbolic checker verifies each answer. gpt-oss-120b 47.5%

Data audit

All 11 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access Public tasks, plus a private test set.
  • Pass Confidence Checked against a primary source.
Scores 5 pass
  • Pass Independent runs Scores from the official leaderboard (46 entries).
  • Pass Models scored 17 models we track, plus 29 other entries on its leaderboard.
  • Pass Newest score The newest is from October 2026, for GPT-6.1 Terra.
  • Pass Frontier models 4 of the top 10 have a score, led by GPT-6 Astra (#4).
  • Pass Headroom The best is 51.3%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on LatchBio: 46 rows for models we track at the last read.
  • Pass Last read November 21st, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check