Skip to content

Benchmark

HELM AIR-Bench: AIRBench 2024 - Fundamental Rights

Safety benchmarks examine risks in model behaviour, including how models respond to unsafe requests. Scores are percentages. Higher is better.

Models measured
70
Scores recorded
75
Sources
1
IQ weight
1.68

Score distribution

77.9%
Mean
86.7%
Median
24.3%
Std dev
40%
10th percentile
100%
90th percentile
5.3%
Min
5.3%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.68 for chat models, measured over 70 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude 3 Haiku· Stanford HELM specialized suites data docs100%
2Claude 3 Opus· Stanford HELM specialized suites data docs100%
3Claude 3 Sonnet· Stanford HELM specialized suites data docs100%
4Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs100%
5Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs100%
6Claude Haiku 4.5· Stanford HELM specialized suites data docs100%
7Claude Sonnet 4· Stanford HELM specialized suites data docs100%
8Claude Sonnet 4.5· Stanford HELM specialized suites data docs100%
9GPT 5.1· Stanford HELM specialized suites data docs100%
10GPT OSS 120B· Stanford HELM specialized suites data docs100%
11GPT OSS 20b· Stanford HELM specialized suites data docs100%
12GPT-5 mini· Stanford HELM specialized suites data docs100%
13GPT-5 nano· Stanford HELM specialized suites data docs100%
14Qwen3 Next 80B A3Bthinking· Stanford HELM specialized suites data docs100%
15GPT 5· Stanford HELM specialized suites data docs98.7%
16Claude Sonnet 3.7· Stanford HELM specialized suites data docs97.3%
17Granite 3.3 8b Instruct· Stanford HELM specialized suites data docs97.3%
18Llama3 8B Chat· Stanford HELM specialized suites data docs97.3%
19o4 Mini· Stanford HELM specialized suites data docs97.3%
20OpenAI o3· Stanford HELM specialized suites data docs97.3%
21Claude Opus 4· Stanford HELM specialized suites data docs96.7%
22Gemini 3 Pro Preview· Stanford HELM specialized suites data docs96%
23o1· Stanford HELM specialized suites data docs96%
24Granite 4.0 H Small· Stanford HELM specialized suites data docs94.7%
25o3 Mini· Stanford HELM specialized suites data docs94.7%
26Qwen3 235B A22B Instruct 2507· Stanford HELM specialized suites data docs94.7%
27GPT-4 Turbo· Stanford HELM specialized suites data docs93.3%
28GPT-4.5 Preview· Stanford HELM specialized suites data docs93.3%
29Palmyra X5· Stanford HELM specialized suites data docs93.3%
30Kimi K2 Instruct· Stanford HELM specialized suites data docs92%
31Gemini 2.5 Pro Preview 03-25· Stanford HELM specialized suites data docs89.3%
32gemini-2.5-flash-preview-04-17· Stanford HELM specialized suites data docs89.3%
33Gemini 2.5 Flash-Lite· Stanford HELM specialized suites data docs88%
34GPT 4.1· Stanford HELM specialized suites data docs86.7%
35GPT-4.1 mini· Stanford HELM specialized suites data docs86.7%
36GPT-4.1 nano· Stanford HELM specialized suites data docs86.7%
37Llama 3.1 8B Instruct Turbo· Stanford HELM specialized suites data docs86.7%
38Qwen2 72B Instruct· Stanford HELM specialized suites data docs86.7%
39GPT 4osetting not named by the source· Stanford HELM specialized suites data docs86%
40GPT-4o mini· Stanford HELM specialized suites data docs84.7%
41Llama 4 Maverick· Stanford HELM specialized suites data docs84.7%
42Gemini 2.0 Flash-Lite Preview· Stanford HELM specialized suites data docs84%
43GPT 4· Stanford HELM specialized suites data docs84%
44Qwen2.5 72B Instruct Turbo· Stanford HELM specialized suites data docs84%
45gemini-2.0-pro-exp-02-05· Stanford HELM specialized suites data docs83.3%
46GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs82.7%
47GLM 4.5 Air· Stanford HELM specialized suites data docs81.3%
48GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs81.3%
49Gemini 2.0 Flash· Stanford HELM specialized suites data docs80%
50Yi 34B Chat· Stanford HELM specialized suites data docs77.3%

Showing 1–50 of 75 results.

Data credits: Stanford HELM (Stanford CRFM); scenario scores from Capabilities, Safety, Long Context, MedHELM, AIR-Bench, MMLU and Finance