Skip to content

Benchmark

HELM AIR-Bench: AIRBench 2024 - Economic Harm

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
0.97

Score distribution

66.2%
Mean
67.3%
Median
26.6%
Std dev
26.3%
10th percentile
97.3%
90th percentile
6.3%
Min
6.3%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.97 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 docs54%
52Yi 34B Chat· Stanford HELM specialized suites data docs53.7%
53GPT 4· Stanford HELM specialized suites data docs51.7%
54Llama 4 Scout· Stanford HELM specialized suites data docs50.7%
55Qwen2.5 7B Instruct Turbo· Stanford HELM specialized suites data docs45.7%
56Grok 3 Mini· Stanford HELM specialized suites data docs44%
57DeepSeek LLM 67B Chat· Stanford HELM specialized suites data docs43.3%
58GPT 4osetting not named by the source· Stanford HELM specialized suites data docs41.3%
59Mixtral 8x22B Instruct v0.1· Stanford HELM specialized suites data docs40.7%
60Qwen1.5-72B-Chat· Stanford HELM specialized suites data docs39.3%
61GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs39%
62O1 Mini· Stanford HELM specialized suites data docs38.3%
63DeepSeek R1· Stanford HELM specialized suites data docs37.7%
64Grok 3· Stanford HELM specialized suites data docs35%
65Meta Llama 3.1 70B Instruct Turbo· Stanford HELM specialized suites data docs34.7%
66GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs30.3%
67Mixtral 8x7B Instruct v0.1· Stanford HELM specialized suites data docs28.3%
68Grok 4· Stanford HELM specialized suites data docs26.3%
69DeepSeek V3· Stanford HELM specialized suites data docs25.3%
70Mistral Large2407· Stanford HELM specialized suites data docs20.7%
71Command R+· Stanford HELM specialized suites data docs19.3%
72Mistral Small· Stanford HELM specialized suites data docs17.7%
73Mistral (7B) Instruct v0.3· Stanford HELM specialized suites data docs14.7%
74Command R Plus· Stanford HELM specialized suites data docs13.3%
75DBRX-instruct· Stanford HELM specialized suites data docs6.3%

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