Skip to content

Benchmark

HELM MMLU: Professional Psychology - EM

Scores are percentages. Higher is better.

Models measured
64
Scores recorded
68
Sources
1
IQ weight
2.24

Score distribution

77.8%
Mean
81.2%
Median
12%
Std dev
64.1%
10th percentile
89.1%
90th percentile
23.2%
Min
23.2%92.2%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.24 for chat models, measured over 64 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 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs92.2%
2Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs92.2%
3gemini-1.5-pro-002· Stanford HELM specialized suites data docs91.2%
4GPT 4osetting not named by the source· Stanford HELM specialized suites data docs90.5%
5Claude 3 Opus· Stanford HELM specialized suites data docs90.4%
6GPT 4osetting not named by the source· Stanford HELM specialized suites data docs89.9%
7gemini-1.5-pro-001· Stanford HELM specialized suites data docs89.4%
8GPT 4· Stanford HELM specialized suites data docs89.1%
9DeepSeek V3· Stanford HELM specialized suites data docs88.7%
10GPT 4 1106 Preview· Stanford HELM specialized suites data docs88.7%
11Qwen2 72B Instruct· Stanford HELM specialized suites data docs88.6%
12Gemini 2.0 Flash Experimental· Stanford HELM specialized suites data docs87.6%
13GPT-4 Turbo· Stanford HELM specialized suites data docs87.3%
14Meta Llama 3 70b· Stanford HELM specialized suites data docs87.1%
15Nova Prov1· Stanford HELM specialized suites data docs86.4%
16Qwen2.5 72B Instruct Turbo· Stanford HELM specialized suites data docs86.4%
17Llama 3.1 405B Instruct Turbo· Stanford HELM specialized suites data docs86.1%
18Mistral Large2407· Stanford HELM specialized suites data docs86.1%
19text-unicorn· Stanford HELM specialized suites data docs85.8%
20Meta Llama 3.1 70B Instruct Turbo· Stanford HELM specialized suites data docs84.6%
21Llama 3.3 70B Instruct Turbo· Stanford HELM specialized suites data docs84.5%
22mixtral-8x22b· Stanford HELM specialized suites data docs84.5%
23Llama-3.2-90B-Vision-Instruct-Turbo· Stanford HELM specialized suites data docs84.3%
24Jamba 1.5 Large· Stanford HELM specialized suites data docs84.2%
25gemma2-27b· Stanford HELM specialized suites data docs84%
26Phi-3 Medium 4K Instruct· Stanford HELM specialized suites data docs83.5%
27Phi-3 Small 8K Instruct· Stanford HELM specialized suites data docs83.5%
28Yi-34B· Stanford HELM specialized suites data docs83.5%
29GPT-4o mini· Stanford HELM specialized suites data docs83.3%
30gemini-1.5-flash-001· Stanford HELM specialized suites data docs82.8%
31Claude Haiku 3.5· Stanford HELM specialized suites data docs82.5%
32Qwen1.5-110B-Chat· Stanford HELM specialized suites data docs82%
33Claude 3 Sonnet· Stanford HELM specialized suites data docs81.4%
34Nova Litev1· Stanford HELM specialized suites data docs81.2%
35DeepSeek LLM 67B Chat· Stanford HELM specialized suites data docs80.9%
36Mistral Largesetting not named by the source· Stanford HELM specialized suites data docs80.9%
37Qwen1.5-72B· Stanford HELM specialized suites data docs80.9%
38gemini-1.5-flash-002· Stanford HELM specialized suites data docs80.6%
39Claude 3 Haiku· Stanford HELM specialized suites data docs80.2%
40DBRX-instruct· Stanford HELM specialized suites data docs80.1%
41Claude 2.1· Stanford HELM specialized suites data docs79.7%
42gemma-2-9b· Stanford HELM specialized suites data docs78.8%
43Mistral Small· Stanford HELM specialized suites data docs76.8%
44Jamba 1.5 Mini· Stanford HELM specialized suites data docs76%
45Llama 2 70b· Stanford HELM specialized suites data docs76%
46Qwen2.5 7B Instruct Turbo· Stanford HELM specialized suites data docs75.7%
47Nova Microv1· Stanford HELM specialized suites data docs73.9%
48Command R Plus· Stanford HELM specialized suites data docs73.5%
49GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs73.2%
50Claude Instant 1.2· Stanford HELM specialized suites data docs72.4%

Showing 1–50 of 68 results.

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