HELM AIR-Bench 2024 - #39-40.34: Characterization of identity - Predicted Personality: leaderboard
HELM AIR-Bench 2024 (v1.19.0) refusal rate for level-4 risk category "#39-40.34: Characterization of identity - Predicted Personality" (level-3 "#39-40: Discrimination/Protected Characteristics Combinations", level-2 "Discrimination/Bias"). Only 30 prompts, so scores move in coarse steps. A GPT-4o judge scores each response from compliance (0) to refusal (1); shown as 0-100, higher means more refusals (safer).
Metric: Refusal Rate (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 87 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 100 | #105 |
| 2 | Claude Sonnet 4.5 | 100 | #148 |
| 3 | GPT-4.1 Mini | 100 | #363 |
| 4 | GPT-OSS-120B | 100 | #346 |
| 5 | GPT-4.1 Nano | 100 | #730 |
| 6 | GPT-5 Nano | 100 | #439 |
| 7 | DeepSeek R1 | 100 | #262 |
| 8 | Llama 4 Scout Instruct | 100 | #547 |
| 9 | GPT-OSS-20B | 100 | #523 |
| 10 | GPT-4o Mini (2024-07-18) | 100 | #568 |
| 11 | DeepSeek V3 | 100 | #329 |
| 12 | GPT-4o (2024-08-06) | 100 | #343 |
| 13 | Llama 3 8B Instruct | 100 | #1116 |
| 14 | GPT-5 Mini (2025-08-07) | 100 | #175 |
| 15 | Claude Haiku 4.5 (20251001) | 100 | #264 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-air-bench-2024-39-40-34-characterization-of-identity-predicted-personality · How It Works · Data refreshed daily, snapshot 2026-11-05.