Benchmark Almanac

VIALS 2

GPT-6 Astra 49.9% leads since October 2026
Not in model cards

Models interpret 161 professional life-science visual artifacts. The score is judged accuracy averaged across three rollouts.

The official changelog requires reruns after three task changes. Values are transcribed from the current chart and hash-checked. Unlabeled confidence intervals are not digitized.

Released
October 2026
Kind
Genomics and life sciences
Level
Expert
Scored
Graded by a model
Answers
Short answer
Tasks
All public
Strongest model tested
GPT-6 Astra #3 on the capability index
Last model added
October 2026
Scores refreshed
November 12th, 2026
14
Models scored on its own leaderboard

GPT-6 Astra leads with 49.9%

Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best.

14 models. Scroll sideways for the rest.

  1. 49.9 GPT-6 Astra
  2. 49.1 GPT-6.1 Sol
  3. 47.2 Gemini 4 Argon
  4. 41.0 Claude Sonnet 5.5
  5. 40.2 Claude Opus 5.5
  6. 31.5 Muse Wonder
  7. 29.4 Gemini 3.8 Flash
  8. 26.5 GPT-5.6 Sol
  9. 24.4 Claude Opus 5
  10. 22.6 Muse Spark 1.2
  11. 20.9 Kimi K3
  12. 18.2 Grok 4.7
  13. 12.6 DeepSeek V4.1 Flash
  14. 10.8 MiniMax-M3
  1. 1 GPT-6 Astra 49.9%
  2. 2 GPT-6.1 Sol 49.1%
  3. 3 Gemini 4 Argon 47.2%
  4. 4 Claude Sonnet 5.5 41%
  5. 5 Claude Opus 5.5 40.2%
  6. 6 Muse Wonder 31.5%
  7. 7 Gemini 3.8 Flash 29.4%
  8. 8 GPT-5.6 Sol 26.5%
  9. 9 Claude Opus 5 24.4%
  10. 10 Muse Spark 1.2 22.6%
  11. 11 Kimi K3 20.9%
  12. 12 Grok 4.7 18.2%
  13. 13 DeepSeek V4.1 Flash 12.6%
  14. 14 MiniMax-M3 10.8%

Source: Official VIALS 2 leaderboard

#ModelScoreSettingDateSource
1 GPT-6 Astra 49.9% Sep 2026 Official leaderboard
2 GPT-6.1 Sol 49.1% Sep 2026 Official leaderboard
3 Gemini 4 Argon 47.2% Sep 2026 Official leaderboard
4 Claude Sonnet 5.5 41% Sep 2026 Official leaderboard
5 Claude Opus 5.5 40.2% Sep 2026 Official leaderboard
6 Muse Wonder 31.5% Oct 2026 Official leaderboard
7 Gemini 3.8 Flash 29.4% Sep 2026 Official leaderboard
8 GPT-5.6 Sol 26.5% Jul 2026 Official leaderboard
9 Claude Opus 5 24.4% Jul 2026 Official leaderboard
10 Muse Spark 1.2 22.6% Aug 2026 Official leaderboard
11 Kimi K3 20.9% Jul 2026 Official leaderboard
12 Grok 4.7 18.2% Sep 2026 Official leaderboard
13 DeepSeek V4.1 Flash 12.6% Sep 2026 Official leaderboard
14 MiniMax-M3 10.8% Jun 2026 Official leaderboard

Number one over time

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

OpenAI
Own leaderboard 1 leader
GPT-6 Astra 49.9%

Source: Official VIALS 2 leaderboard

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 Best so far
0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) GPT-6 Astra Gemini 3.8 Flash MiniMax-M3 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) GPT-6 Astra Gemini 3.8 Flash

Source: Official VIALS 2 leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

Versions

Replaces VIALS

More in Biology and chemistry

All 108 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%
  • 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%
  • MedChemBench (OpenAI) In use Models answer chemistry questions from an OpenAI internal set, grouped with OpenAI's life science research evaluations. OpenAI has not published the task format. GPT-Rosalind-6 49.6%
  • 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%