Skip to content

Benchmark

HELM AIR-Bench: AIRBench 2024 - #17: Adult content

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
0.00

Score distribution

65.6%
Mean
65.2%
Median
14.2%
Std dev
46.4%
10th percentile
83.3%
90th percentile
27.5%
Min
27.5%96.4%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.00 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
51Yi 34B Chat· Stanford HELM specialized suites data docs60.1%
52DeepSeek LLM 67B Chat· Stanford HELM specialized suites data docs59.4%
53OpenAI o3· Stanford HELM specialized suites data docs58.7%
54gemini-1.5-flash-002· Stanford HELM specialized suites data docs57.2%
55Llama 3.1 405B Instruct Turbo· Stanford HELM specialized suites data docs57.2%
56Meta Llama 3.1 70B Instruct Turbo· Stanford HELM specialized suites data docs57.2%
57DeepSeek R1· Stanford HELM specialized suites data docs55.1%
58Gemini 2.5 Flash-Lite· Stanford HELM specialized suites data docs53.6%
59Qwen1.5-72B-Chat· Stanford HELM specialized suites data docs52.9%
60gemini-1.5-pro-002· Stanford HELM specialized suites data docs52.2%
61o4 Mini· Stanford HELM specialized suites data docs51.4%
62GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs49.3%
63Command R Plus· Stanford HELM specialized suites data docs48.6%
64GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs48.6%
65o1· Stanford HELM specialized suites data docs48.6%
66Gemini 2.5 Pro Preview 03-25· Stanford HELM specialized suites data docs46.4%
67gemini-2.5-flash-preview-04-17· Stanford HELM specialized suites data docs46.4%
68Grok 3· Stanford HELM specialized suites data docs46.4%
69o3 Mini· Stanford HELM specialized suites data docs46.4%
70GLM 4.5 Air· Stanford HELM specialized suites data docs45.7%
71Kimi K2 Instruct· Stanford HELM specialized suites data docs44.2%
72Command R+· Stanford HELM specialized suites data docs42.8%
73Grok 3 Mini· Stanford HELM specialized suites data docs40.6%
74Gemini 3 Pro Preview· Stanford HELM specialized suites data docs34.1%
75Grok 4· Stanford HELM specialized suites data docs27.5%

Showing 51–75 of 75 results.

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