SecureBio BioTIER Refuse: leaderboard

SecureBio BioTIER Refuse — 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: Hazardous prompts refused (%). Source: https://securebio.org/benchmarks/details/

Metric: Hazardous prompts refused (%). Source: securebio.org. Saturation forecast: Around February 2027. 77 models tracked. Updated 16 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Claude Opus 5.5Max99.1#5
2Claude Sonnet 4.5 [16k]SecureBio run: native-setting=Claude Sonnet 4.5 [16k]98.2
3Claude Haiku 5.5Max97.9#70
4Claude Sonnet 5Max96.1#111
5Claude Opus 4.8Max95.9#58
6Claude Sonnet 5.5Max95.7#19
7Muse Spark OpenxHigh95.5#81
8Claude Sonnet 4.6Max95.2#165
9Claude Opus 5Max94.9#20
10Claude Opus 4.7Max94.9#78
11Claude Opus 4.6Max94.4#106
12Claude Opus 4.5 [16k]SecureBio run: native-setting=Claude Opus 4.5 [16k]92.6
13Gemini 4 ArgonHigh87.8#12
14Gemini 3.5 FlashHigh81.7#100
15Gemini 3.6 FlashHigh81.1#75

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon.

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