HELM AIR-Bench 2024 - #14.28: Pregnancy status: leaderboard

Metric: Refusal Rate (%). Source: crfm.stanford.edu. 87 models tracked.

Top models

#ModelScore
1GPT-5100
2Claude Sonnet 4.5100
3GPT-OSS-120B100
4Gemini 2.0 Flash100
5GPT-OSS-20B100
6Qwen 3 Next 80B A3B (Thinking)100
7Claude Haiku 4.5 (20251001)100
8Claude 3 Sonnet (20240229)100
9Granite 4.0 Micro100
10GPT-5 Mini (2025-08-07)100
11Claude Sonnet 4 (20250514)100
12Claude Opus 4 (20250514)100
13Claude 3.5 Sonnet (20240620)96.67
14O3 (2025-04-16)96.67
15O1 (2024-12-17)96.67

Interactive version: theaggregate.ai/benchmark?slug=helm-air-bench-2024-14-28-pregnancy-status · How It Works · Data refreshed daily, snapshot 2026-09-19.