Skip to content

Benchmark

HELM AIR-Bench: AIRBench 2024 - #28: High-Risk Financial Activities

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
1.46

Score distribution

76.4%
Mean
83.3%
Median
24.4%
Std dev
33.3%
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 1.46 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 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs100%
5Claude Haiku 4.5· Stanford HELM specialized suites data docs100%
6Claude Sonnet 4.5· Stanford HELM specialized suites data docs100%
7GPT OSS 120B· Stanford HELM specialized suites data docs100%
8GPT OSS 20b· Stanford HELM specialized suites data docs100%
9Granite 3.3 8b Instruct· Stanford HELM specialized suites data docs100%
10Llama3 8B Chat· Stanford HELM specialized suites data docs100%
11OpenAI o3· Stanford HELM specialized suites data docs100%
12Qwen3 Next 80B A3Bthinking· Stanford HELM specialized suites data docs100%
13Claude 3 Opus· Stanford HELM specialized suites data docs96.7%
14Claude Opus 4· Stanford HELM specialized suites data docs96.7%
15Claude Sonnet 4· Stanford HELM specialized suites data docs96.7%
16GPT-5 nano· Stanford HELM specialized suites data docs96.7%
17Llama 3.1 8B Instruct Turbo· Stanford HELM specialized suites data docs96.7%
18o1· Stanford HELM specialized suites data docs96.7%
19o4 Mini· Stanford HELM specialized suites data docs96.7%
20Palmyra X5· Stanford HELM specialized suites data docs96.7%
21Qwen3 235B A22B Instruct 2507· Stanford HELM specialized suites data docs96.7%
22GPT-4.1 nano· Stanford HELM specialized suites data docs95%
23Gemini 2.0 Flash· Stanford HELM specialized suites data docs93.3%
24gemini-2.0-pro-exp-02-05· Stanford HELM specialized suites data docs93.3%
25GPT 5.1· Stanford HELM specialized suites data docs93.3%
26GPT-4.5 Preview· Stanford HELM specialized suites data docs93.3%
27GPT-5 mini· Stanford HELM specialized suites data docs93.3%
28Llama 4 Maverick· Stanford HELM specialized suites data docs93.3%
29Claude Sonnet 3.7· Stanford HELM specialized suites data docs90%
30Gemini 2.0 Flash-Lite Preview· Stanford HELM specialized suites data docs90%
31Gemini 2.5 Flash-Lite· Stanford HELM specialized suites data docs90%
32Gemini 3 Pro Preview· Stanford HELM specialized suites data docs90%
33GPT-4 Turbo· Stanford HELM specialized suites data docs90%
34Granite 4.0 H Small· Stanford HELM specialized suites data docs90%
35GPT 5· Stanford HELM specialized suites data docs86.7%
36GPT-4o mini· Stanford HELM specialized suites data docs86.7%
37o3 Mini· Stanford HELM specialized suites data docs86.7%
38GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs83.3%
39gemini-1.5-flash-002· Stanford HELM specialized suites data docs80%
40GLM 4.5 Air· Stanford HELM specialized suites data docs80%
41GPT 4.1· Stanford HELM specialized suites data docs80%
42GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs80%
43Kimi K2 Instruct· Stanford HELM specialized suites data docs80%
44Gemini 2.5 Pro Preview 03-25· Stanford HELM specialized suites data docs76.7%
45gemini-1.5-pro-002· Stanford HELM specialized suites data docs76.7%
46gemini-2.5-flash-preview-04-17· Stanford HELM specialized suites data docs76.7%
47GPT-4.1 mini· Stanford HELM specialized suites data docs76.7%
48Qwen2 72B Instruct· Stanford HELM specialized suites data docs76.7%
49GPT 4· Stanford HELM specialized suites data docs73.3%
50Llama 4 Scout· Stanford HELM specialized suites data docs73.3%

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