Skip to content

Benchmark

HELM AIR-Bench: AIRBench 2024 - #29: Unfair Market Practices

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
0.78

Score distribution

55.3%
Mean
55%
Median
34.1%
Std dev
10%
10th percentile
100%
90th percentile
3.3%
Min
3.3%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.78 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 Sonnet· Stanford HELM specialized suites data docs100%
3Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs100%
4Claude Haiku 4.5· Stanford HELM specialized suites data docs100%
5GPT 5· Stanford HELM specialized suites data docs100%
6GPT OSS 120B· Stanford HELM specialized suites data docs100%
7GPT-5 mini· Stanford HELM specialized suites data docs100%
8GPT-5 nano· Stanford HELM specialized suites data docs100%
9o1· Stanford HELM specialized suites data docs100%
10OpenAI o3· Stanford HELM specialized suites data docs100%
11Palmyra X5· Stanford HELM specialized suites data docs100%
12Qwen3 Next 80B A3Bthinking· Stanford HELM specialized suites data docs98.3%
13Claude 3 Opus· Stanford HELM specialized suites data docs96.7%
14Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs96.7%
15Claude Sonnet 4.5· Stanford HELM specialized suites data docs96.7%
16GPT 5.1· Stanford HELM specialized suites data docs96.7%
17GPT OSS 20b· Stanford HELM specialized suites data docs96.7%
18Claude Sonnet 4· Stanford HELM specialized suites data docs93.3%
19Kimi K2 Instruct· Stanford HELM specialized suites data docs93.3%
20Qwen3 235B A22B Instruct 2507· Stanford HELM specialized suites data docs91.7%
21GPT-4.5 Preview· Stanford HELM specialized suites data docs90%
22Claude Opus 4· Stanford HELM specialized suites data docs88.3%
23o4 Mini· Stanford HELM specialized suites data docs86.7%
24Claude Sonnet 3.7· Stanford HELM specialized suites data docs85%
25Llama 4 Maverick· Stanford HELM specialized suites data docs85%
26Granite 3.3 8b Instruct· Stanford HELM specialized suites data docs80%
27Llama3 8B Chat· Stanford HELM specialized suites data docs80%
28Gemini 3 Pro Preview· Stanford HELM specialized suites data docs73.3%
29o3 Mini· Stanford HELM specialized suites data docs73.3%
30Llama 3.1 8B Instruct Turbo· Stanford HELM specialized suites data docs68.3%
31GPT 4.1· Stanford HELM specialized suites data docs65%
32Gemini 2.5 Pro Preview 03-25· Stanford HELM specialized suites data docs63.3%
33gemini-2.0-pro-exp-02-05· Stanford HELM specialized suites data docs63.3%
34GPT-4 Turbo· Stanford HELM specialized suites data docs63.3%
35gemini-2.5-flash-preview-04-17· Stanford HELM specialized suites data docs60%
36Granite 4.0 H Small· Stanford HELM specialized suites data docs58.3%
37gemini-1.5-pro-002· Stanford HELM specialized suites data docs56.7%
38Gemini 2.0 Flash-Lite Preview· Stanford HELM specialized suites data docs55%
39GLM 4.5 Air· Stanford HELM specialized suites data docs51.7%
40Llama 3.1 405B Instruct Turbo· Stanford HELM specialized suites data docs50%
41Gemini 2.0 Flash· Stanford HELM specialized suites data docs43.3%
42Qwen2 72B Instruct· Stanford HELM specialized suites data docs43.3%
43GPT-4.1 mini· Stanford HELM specialized suites data docs41.7%
44GPT-4.1 nano· Stanford HELM specialized suites data docs41.7%
45O1 Mini· Stanford HELM specialized suites data docs41.7%
46GPT 4osetting not named by the source· Stanford HELM specialized suites data docs40%
47GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs36.7%
48Gemini 2.5 Flash-Lite· Stanford HELM specialized suites data docs33.3%
49gemini-1.5-flash-002· Stanford HELM specialized suites data docs33.3%
50Qwen2.5 72B Instruct Turbo· Stanford HELM specialized suites data docs33.3%

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