Skip to content

Benchmark

HELM AIR-Bench: AIRBench 2024 - Sexual 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
0.48

Score distribution

55.8%
Mean
50.7%
Median
16.3%
Std dev
38.5%
10th percentile
80.2%
90th percentile
24.7%
Min
24.7%95.1%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.48 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
51O1 Mini· Stanford HELM specialized suites data docs45.8%
52Gemini 2.5 Pro Preview 03-25· Stanford HELM specialized suites data docs44.8%
53Llama 3.1 405B Instruct Turbo· Stanford HELM specialized suites data docs44.8%
54DeepSeek R1· Stanford HELM specialized suites data docs44.4%
55Mixtral 8x22B Instruct v0.1· Stanford HELM specialized suites data docs44.4%
56Mixtral 8x7B Instruct v0.1· Stanford HELM specialized suites data docs43.8%
57Meta Llama 3.1 70B Instruct Turbo· Stanford HELM specialized suites data docs43.4%
58DeepSeek V3· Stanford HELM specialized suites data docs42.7%
59GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs42.4%
60DeepSeek LLM 67B Chat· Stanford HELM specialized suites data docs42%
61Mistral (7B) Instruct v0.3· Stanford HELM specialized suites data docs42%
62gemini-2.5-flash-preview-04-17· Stanford HELM specialized suites data docs40.3%
63Mistral Small· Stanford HELM specialized suites data docs39.9%
64o1· Stanford HELM specialized suites data docs39.6%
65Qwen1.5-72B-Chat· Stanford HELM specialized suites data docs39.6%
66GLM 4.5 Air· Stanford HELM specialized suites data docs39.2%
67o3 Mini· Stanford HELM specialized suites data docs39.2%
68DBRX-instruct· Stanford HELM specialized suites data docs38.5%
69GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs36.5%
70Grok 3 Mini· Stanford HELM specialized suites data docs36.5%
71Gemini 3 Pro Preview· Stanford HELM specialized suites data docs35.8%
72Grok 3· Stanford HELM specialized suites data docs34.7%
73Command R Plus· Stanford HELM specialized suites data docs32.6%
74Command R+· Stanford HELM specialized suites data docs29.9%
75Grok 4· Stanford HELM specialized suites data docs24.7%

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