Skip to content

Benchmark

HELM MMLU: Professional Medicine - EM

Scores are percentages. Higher is better.

Models measured
64
Scores recorded
68
Sources
1
IQ weight
2.10

Score distribution

78.8%
Mean
82.4%
Median
13.9%
Std dev
58.1%
10th percentile
93.4%
90th percentile
44.9%
Min
44.9%96%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.10 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
1GPT 4osetting not named by the source· Stanford HELM specialized suites data docs96%
2Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs94.9%
3DeepSeek V3· Stanford HELM specialized suites data docs94.9%
4GPT 4osetting not named by the source· Stanford HELM specialized suites data docs94.9%
5Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs94.5%
6GPT 4 1106 Preview· Stanford HELM specialized suites data docs94.5%
7gemini-1.5-pro-002· Stanford HELM specialized suites data docs93.8%
8Claude 3 Opus· Stanford HELM specialized suites data docs93.4%
9GPT 4· Stanford HELM specialized suites data docs93.4%
10Llama 3.1 405B Instruct Turbo· Stanford HELM specialized suites data docs93.4%
11Llama-3.2-90B-Vision-Instruct-Turbo· Stanford HELM specialized suites data docs93.4%
12Meta Llama 3.1 70B Instruct Turbo· Stanford HELM specialized suites data docs93%
13GPT-4 Turbo· Stanford HELM specialized suites data docs91.5%
14Mistral Large2407· Stanford HELM specialized suites data docs90.8%
15Qwen2.5 72B Instruct Turbo· Stanford HELM specialized suites data docs90.8%
16Nova Prov1· Stanford HELM specialized suites data docs90.4%
17Gemini 2.0 Flash Experimental· Stanford HELM specialized suites data docs89.7%
18Llama 3.3 70B Instruct Turbo· Stanford HELM specialized suites data docs89.3%
19gemini-1.5-pro-001· Stanford HELM specialized suites data docs88.6%
20Qwen2 72B Instruct· Stanford HELM specialized suites data docs88.6%
21Meta Llama 3 70b· Stanford HELM specialized suites data docs88.2%
22Jamba 1.5 Large· Stanford HELM specialized suites data docs87.5%
23mixtral-8x22b· Stanford HELM specialized suites data docs87.1%
24Claude 3 Sonnet· Stanford HELM specialized suites data docs86.8%
25Phi-3 Medium 4K Instruct· Stanford HELM specialized suites data docs86.8%
26GPT-4o mini· Stanford HELM specialized suites data docs86.4%
27Claude Haiku 3.5· Stanford HELM specialized suites data docs86%
28Qwen1.5-110B-Chat· Stanford HELM specialized suites data docs86%
29Claude 3 Haiku· Stanford HELM specialized suites data docs85.7%
30text-unicorn· Stanford HELM specialized suites data docs84.9%
31gemma2-27b· Stanford HELM specialized suites data docs83.5%
32Nova Litev1· Stanford HELM specialized suites data docs83.1%
33Phi-3 Small 8K Instruct· Stanford HELM specialized suites data docs82.4%
34Qwen1.5-72B· Stanford HELM specialized suites data docs82.4%
35DBRX-instruct· Stanford HELM specialized suites data docs82%
36Yi-34B· Stanford HELM specialized suites data docs82%
37gemini-1.5-flash-001· Stanford HELM specialized suites data docs81.3%
38Nova Microv1· Stanford HELM specialized suites data docs80.9%
39Claude 2.1· Stanford HELM specialized suites data docs79.8%
40GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs79.4%
41gemma-2-9b· Stanford HELM specialized suites data docs77.9%
42Jamba 1.5 Mini· Stanford HELM specialized suites data docs77.9%
43Qwen2.5 7B Instruct Turbo· Stanford HELM specialized suites data docs77.9%
44GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs77.2%
45DeepSeek LLM 67B Chat· Stanford HELM specialized suites data docs76.5%
46Open Mistral Nemo· Stanford HELM specialized suites data docs76.1%
47Claude Instant 1.2· Stanford HELM specialized suites data docs75%
48Llama 2 70b· Stanford HELM specialized suites data docs75%
49Mistral Small· Stanford HELM specialized suites data docs74.6%
50Command R Plus· Stanford HELM specialized suites data docs73.2%

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