SecureBio VCT v2: leaderboard

SecureBio VCT v2 — Native SecureBio benchmark panel; exact named population. Accuracy includes native refusal/failure conventions and source-reported runs. Raw fractions converted only according to this panel: ABLEFull already uses percent, baseline-only and other capability panels use fractions. Metric: Accuracy (%). Source: https://securebio.org/benchmarks/details/

Metric: Accuracy (%). Source: securebio.org. Saturation forecast: Around 2029. 27 models tracked. Updated 16 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Claude Opus 5.5Max60.54#5
2GPT-6 AstraMax58#6
3Gemini 4 ArgonHigh57.35#12
4Claude Sonnet 5.5Max52.18#19
5GPT-Rosalind-5.5xHigh50.29#46
6GPT-5.6 SolMax49.45#30
7GPT-6 SolMax49.35#26
8Gemini 3.7 FlashHigh48.63#34
9Gemini 3.5 FlashHigh46.51#100
10Kimi K3.1 ProMax46.24#29
11Claude Opus 4.8Max45.87#58
12GPT-5.5 ProxHigh45.34#35
13GPT-5.5xHigh45.27#52
14Gemini 3.8 FlashHigh45.23#36
15Gemini 3.1 Pro (Preview)High43.64#93

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #10 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=securebio-vct-v2 · How It Works · Data refreshed daily, snapshot 2026-11-16.