Benchmark Almanac

VCT-v2

Claude Opus 5.5 60.5% leads since September 2026
Only in system cards

Models answer an updated set of expert virology troubleshooting questions from SecureBio. Scored as refusal-corrected accuracy.

279 text and image questions with a private holdout. 35 PhD virologists give the expert baseline, reusing 216 original VCT expert answers rescored.

Made by
SecureBio
Released
September 2026
Kind
Lab protocols
Level
Expert
Scored
Checked automatically
Answers
Multiple choice
Tasks
Kept private
Strongest model tested
GPT-6 Astra #3 on the capability index
Last in a system card
September 2026
Last model added
October 2026
Scores refreshed
November 12th, 2026
27
Entries on its own leaderboard, 25 models we track
1
System card reports it, from 1 lab

Claude Opus 5.5 leads with 60.5%

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.

27 entries. Scroll sideways for the rest.

  1. 60.5 Claude Opus 5.5
  2. 58.0 GPT-6 Astra
  3. 57.3 Gemini 4 Argon
  4. 52.2 Claude Sonnet 5.5
  5. 50.3 GPT-5.5 Rosalind
  6. 49.4 GPT-5.6 Sol
  7. 49.3 GPT-6 Sol
  8. 48.6 Gemini 3.7 Flash
  9. 46.5 Gemini 3.5 Flash
  10. 46.2 Kimi K3.1 Pro
  11. 45.9 Claude Opus 4.8
  12. 45.3 GPT-5.5
  13. 45.3 GPT-5.5 Pro
  14. 45.2 Gemini 3.8 Flash
  15. 43.6 Gemini 3.1 Pro (Feb 26' Preview)
  16. 42.0 Grok 4.7
  17. 41.9 GPT-5.6 Terra
  18. 41.8 Kimi K3
  19. 41.2 Grok 4.5
  20. 39.4 Mistral Large 4
  21. 38.5 DeepSeek V4.1 Flash 1017
  22. 37.2 Kimi K2.6
  23. 34.5 Inkling-Small
  24. 31.9 Inkling
  25. 28.0 Qwen 3.8 27B
  26. 26.0 Nemotron 3.5 Super VL
  27. 20.0 Claude Haiku 4.5
  1. 1 Claude Opus 5.5 max 60.5%
  2. 2 GPT-6 Astra max 58%
  3. 3 Gemini 4 Argon high 57.3%
  4. 4 Claude Sonnet 5.5 max 52.2%
  5. 5 GPT-5.5 Rosalind xhigh 50.3%
  6. 6 GPT-5.6 Sol max 49.4%
  7. 7 GPT-6 Sol max 49.3%
  8. 8 Gemini 3.7 Flash high 48.6%
  9. 9 Gemini 3.5 Flash high 46.5%
  10. 10 Kimi K3.1 Pro max 46.2%
  11. 11 Claude Opus 4.8 max 45.9%
  12. 12 GPT-5.5 xhigh 45.3%
  13. 13 GPT-5.5 Pro xhigh 45.3%
  14. 14 Gemini 3.8 Flash high 45.2%
  15. 15 Gemini 3.1 Pro (Feb 26' Preview) high 43.6%
  16. 16 Grok 4.7 xhigh 42%
  17. 17 GPT-5.6 Terra max 41.9%
  18. 18 Kimi K3 max 41.8%
  19. 19 Grok 4.5 41.2%
  20. 20 Mistral Large 4 high 39.4%
  21. 21 DeepSeek V4.1 Flash 1017 max 38.5%
  22. 22 Kimi K2.6 37.2%
  23. 23 Inkling-Small xhigh 34.5%
  24. 24 Inkling xhigh 31.9%
  25. 25 Qwen 3.8 27B xhigh 28%
  26. 26 Nemotron 3.5 Super VL 26%
  27. 27 Claude Haiku 4.5 16k 20%

Source: Official VCT-v2 leaderboard

#EntryScoreSettingDateSource
1 Claude Opus 5.5 60.5% max Sep 2026 Official leaderboard
2 GPT-6 Astra 58% max Sep 2026 Official leaderboard
3 Gemini 4 Argon 57.3% high Sep 2026 Official leaderboard
4 Claude Sonnet 5.5 52.2% max Sep 2026 Official leaderboard
5 GPT-5.5 Rosalind 50.3% xhigh Jun 2026 Official leaderboard
6 GPT-5.6 Sol 49.4% max Jul 2026 Official leaderboard
7 GPT-6 Sol 49.3% max Sep 2026 Official leaderboard
8 Gemini 3.7 Flash 48.6% high Aug 2026 Official leaderboard
9 Gemini 3.5 Flash 46.5% high May 2026 Official leaderboard
10 Kimi K3.1 Pro 46.2% max Oct 2026 Official leaderboard
11 Claude Opus 4.8 45.9% max May 2026 Official leaderboard
12 GPT-5.5 45.3% xhigh Apr 2026 Official leaderboard
13 GPT-5.5 Pro 45.3% xhigh Apr 2026 Official leaderboard
14 Gemini 3.8 Flash 45.2% high Sep 2026 Official leaderboard
15 Gemini 3.1 Pro (Feb 26' Preview) 43.6% high Feb 2026 Official leaderboard
16 Grok 4.7 42% xhigh Sep 2026 Official leaderboard
17 GPT-5.6 Terra 41.9% max Jul 2026 Official leaderboard
18 Kimi K3 41.8% max Jul 2026 Official leaderboard
19 Grok 4.5 41.2% Jul 2026 Official leaderboard
20 Mistral Large 4 39.4% high Oct 2026 Official leaderboard
21 DeepSeek V4.1 Flash 1017 38.5% max Oct 2026 Official leaderboard
22 Kimi K2.6 37.2% Apr 2026 Official leaderboard
23 Inkling-Small 34.5% xhigh Jul 2026 Official leaderboard
24 Inkling 31.9% xhigh Jul 2026 Official leaderboard
25 Qwen 3.8 27B 28% xhigh Aug 2026 Official leaderboard
26 Nemotron 3.5 Super VL 26% Oct 2026 Official leaderboard
27 Claude Haiku 4.5 20% 16k Oct 2025 Official leaderboard

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

3 models have held number one on the official leaderboard since its release in September 2026

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

OpenAI Anthropic
Own leaderboard 3 leaders
Claude Opus 5.5 60.5%

Source: Official VCT-v2 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 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 GPT-5.6 Sol Kimi K2.6 Claude Haiku 4.5 0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) Claude Opus 5.5 GPT-5.6 Sol

Sources: Official VCT-v2 leaderboard and labs’ model cards

Labs report it only in system cards

We read the headline results of 522 launch posts from 52 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
Sep 3, 2026 GPT-6 Astra system card 57.7 SecureBio VCT-v2, refusal-corrected accuracy, pre-release checkpoint (10.1.1.4) Bio and chem risk

One lab claim next to the leaderboard

Each score a lab reported for its own model, beside the same model's best score on the leaderboard. Gaps mostly come from different harnesses, effort levels and task subsets. Every benchmark.

Model Lab says Board says Gap
GPT-6 Astra 57.7% SecureBio VCT-v2, refusal-corrected accuracy, pre-release checkpoint (10.1.1.4) 58% Leaderboard, max −0.3 pts

Versions

We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.

OpenAI xAI Anthropic
VCT-v2 8 leaders
Claude Opus 5.5 60.5%
Virology Capabilities Test (VCT) 5 leaders
Grok 4.6 67.4%
VCT-v2 3 leaders
Claude Opus 5.5 60.5%

Sources: Official leaderboards and labs’ model cards

Replaces Virology Capabilities Test (VCT)

More in Biology and chemistry

All 109 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%