Skip to content

Benchmark

HELM MMLU: Electrical Engineering - EM

Scores are percentages. Higher is better.

Models measured
64
Scores recorded
68
Sources
1
IQ weight
2.28

Score distribution

69.8%
Mean
72.4%
Median
11.3%
Std dev
54.5%
10th percentile
80.7%
90th percentile
29%
Min
29%86.9%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.28 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
1DeepSeek V3· Stanford HELM specialized suites data docs86.9%
2gemini-1.5-pro-002· Stanford HELM specialized suites data docs85.5%
3Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs84.8%
4Gemini 2.0 Flash Experimental· Stanford HELM specialized suites data docs83.4%
5Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs82.8%
6Llama 3.1 405B Instruct Turbo· Stanford HELM specialized suites data docs82.1%
7Claude 3 Opus· Stanford HELM specialized suites data docs81.4%
8GPT 4osetting not named by the source· Stanford HELM specialized suites data docs80.7%
9gemini-1.5-flash-001· Stanford HELM specialized suites data docs80%
10Nova Prov1· Stanford HELM specialized suites data docs80%
11Qwen2.5 72B Instruct Turbo· Stanford HELM specialized suites data docs80%
12GPT 4osetting not named by the source· Stanford HELM specialized suites data docs79.3%
13Jamba 1.5 Large· Stanford HELM specialized suites data docs79.3%
14Mistral Large2407· Stanford HELM specialized suites data docs79.3%
15Qwen2 72B Instruct· Stanford HELM specialized suites data docs79.3%
16GPT 4· Stanford HELM specialized suites data docs78.6%
17Nova Litev1· Stanford HELM specialized suites data docs77.9%
18Qwen1.5-72B· Stanford HELM specialized suites data docs77.9%
19Yi-34B· Stanford HELM specialized suites data docs77.9%
20gemini-1.5-flash-002· Stanford HELM specialized suites data docs77.2%
21GPT 4 1106 Preview· Stanford HELM specialized suites data docs77.2%
22text-unicorn· Stanford HELM specialized suites data docs77.2%
23Meta Llama 3 70b· Stanford HELM specialized suites data docs76.6%
24mixtral-8x22b· Stanford HELM specialized suites data docs76.6%
25Llama-3.2-90B-Vision-Instruct-Turbo· Stanford HELM specialized suites data docs75.9%
26GPT-4 Turbo· Stanford HELM specialized suites data docs75.2%
27Qwen1.5-110B-Chat· Stanford HELM specialized suites data docs75.2%
28gemini-1.5-pro-001· Stanford HELM specialized suites data docs74.5%
29Llama 3.3 70B Instruct Turbo· Stanford HELM specialized suites data docs74.5%
30Meta Llama 3.1 70B Instruct Turbo· Stanford HELM specialized suites data docs74.5%
31gemma2-27b· Stanford HELM specialized suites data docs73.8%
32GPT-4o mini· Stanford HELM specialized suites data docs73.1%
33Claude 2.1· Stanford HELM specialized suites data docs72.4%
34gemma-2-9b· Stanford HELM specialized suites data docs72.4%
35Claude Haiku 3.5· Stanford HELM specialized suites data docs71.7%
36Qwen2.5 7B Instruct Turbo· Stanford HELM specialized suites data docs71.7%
37Command R Plus· Stanford HELM specialized suites data docs71%
38DBRX-instruct· Stanford HELM specialized suites data docs71%
39Claude 3 Sonnet· Stanford HELM specialized suites data docs70.3%
40Claude 3 Haiku· Stanford HELM specialized suites data docs69%
41Phi-3 Small 8K Instruct· Stanford HELM specialized suites data docs69%
42Jamba 1.5 Mini· Stanford HELM specialized suites data docs68.3%
43Nova Microv1· Stanford HELM specialized suites data docs68.3%
44Phi-3 Medium 4K Instruct· Stanford HELM specialized suites data docs68.3%
45Qwen1.5-14B· Stanford HELM specialized suites data docs68.3%
46DeepSeek LLM 67B Chat· Stanford HELM specialized suites data docs66.9%
47GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs66.9%
48Meta Llama 3 8b· Stanford HELM specialized suites data docs66.9%
49Snowflake Arctic Instruct· Stanford HELM specialized suites data docs66.2%
50Yi-6B· Stanford HELM specialized suites data docs66.2%

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