SecureBio BioTIER Permit: leaderboard
SecureBio BioTIER Permit — BioTIER separate benign compliance and hazardous refusal behavior panels, not biological task capability. Native percentages and model/run labels retained; no automatic eligibility/exclusion policy. Metric: Benign prompts answered (%). Source: https://securebio.org/benchmarks/details/
Metric: Benign prompts answered (%). Source: securebio.org. Saturation forecast: Not forecast. 77 models tracked. Updated 16 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | GPT-Rosalind-5.5 | xHigh | 100 | #46 |
| 2 | GPT-4.1 | 99.9 | #500 | |
| 3 | DeepSeek V3.2 | 99.9 | #412 | |
| 4 | Mistral Large 3 | 99.9 | #827 | |
| 5 | Kimi K2 (Thinking) | 99.9 | ||
| 6 | Qwen 3.7 Max | 99.9 | #147 | |
| 7 | Mistral Medium 3.1 | 99.9 | #863 | |
| 8 | Gemini 3.1 Pro (Preview) | High | 99.9 | #93 |
| 9 | Gemini 3.5 Flash | High | 99.9 | #100 |
| 10 | Mistral Large 4 | High | 99.9 | #137 |
| 11 | Kimi K3.1 Pro | Max | 99.9 | #29 |
| 12 | Inkling | xHigh | 99.9 | #281 |
| 13 | Nemotron 3.5 Super VL | 99.9 | #349 | |
| 14 | Qwen3.8 2.4T A95B | xHigh | 99.9 | #83 |
| 15 | GPT-4.1 Mini | 99.8 | #781 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=securebio-biotier-permit · How It Works · Data refreshed daily, snapshot 2026-11-16.