Skip to content

Benchmark

HELM AIR-Bench: AIRBench 2024 - #22: Child Sexual Abuse

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
1.40

Score distribution

74.1%
Mean
80%
Median
21.8%
Std dev
40%
10th percentile
96.7%
90th percentile
6.7%
Min
6.7%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.40 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
51GPT 4osetting not named by the source· Stanford HELM specialized suites data docs66.7%
52GPT-4o mini· Stanford HELM specialized suites data docs66.7%
53Grok 4· Stanford HELM specialized suites data docs66.7%
54Llama 3.1 405B Instruct Turbo· Stanford HELM specialized suites data docs66.7%
55Qwen2.5 72B Instruct Turbo· Stanford HELM specialized suites data docs66.7%
56GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs63.3%
57Qwen2 72B Instruct· Stanford HELM specialized suites data docs63.3%
58DeepSeek R1· Stanford HELM specialized suites data docs60%
59Meta Llama 3.1 70B Instruct Turbo· Stanford HELM specialized suites data docs60%
60Grok 3· Stanford HELM specialized suites data docs58.3%
61Qwen1.5-72B-Chat· Stanford HELM specialized suites data docs56.7%
62Qwen2.5 7B Instruct Turbo· Stanford HELM specialized suites data docs56.7%
63Yi 34B Chat· Stanford HELM specialized suites data docs56.7%
64O1 Mini· Stanford HELM specialized suites data docs53.3%
65DeepSeek V3· Stanford HELM specialized suites data docs50%
66DeepSeek LLM 67B Chat· Stanford HELM specialized suites data docs46.7%
67Mistral Large2407· Stanford HELM specialized suites data docs43.3%
68GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs40%
69Command R Plus· Stanford HELM specialized suites data docs36.7%
70Command R+· Stanford HELM specialized suites data docs36.7%
71Mistral Small· Stanford HELM specialized suites data docs33.3%
72Mixtral 8x22B Instruct v0.1· Stanford HELM specialized suites data docs33.3%
73Mixtral 8x7B Instruct v0.1· Stanford HELM specialized suites data docs26.7%
74DBRX-instruct· Stanford HELM specialized suites data docs6.7%
75Mistral (7B) Instruct v0.3· Stanford HELM specialized suites data docs6.7%

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