Skip to content

Benchmark

HELM MMLU: High School Microeconomics - EM

Scores are percentages. Higher is better.

Models measured
64
Scores recorded
68
Sources
1
IQ weight
1.98

Score distribution

82.4%
Mean
85.7%
Median
12.8%
Std dev
68.9%
10th percentile
94.5%
90th percentile
29.4%
Min
29.4%97.1%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.98 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 docs97.1%
2DeepSeek V3· Stanford HELM specialized suites data docs96.6%
3Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs95.8%
4GPT 4osetting not named by the source· Stanford HELM specialized suites data docs95.8%
5Gemini 2.0 Flash Experimental· Stanford HELM specialized suites data docs95.4%
6Llama 3.1 405B Instruct Turbo· Stanford HELM specialized suites data docs95.4%
7gemini-1.5-pro-002· Stanford HELM specialized suites data docs95%
8Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs94.5%
9GPT 4 1106 Preview· Stanford HELM specialized suites data docs94.5%
10GPT-4 Turbo· Stanford HELM specialized suites data docs94.5%
11Qwen2.5 72B Instruct Turbo· Stanford HELM specialized suites data docs94.5%
12gemini-1.5-pro-001· Stanford HELM specialized suites data docs94.1%
13GPT 4· Stanford HELM specialized suites data docs93.7%
14text-unicorn· Stanford HELM specialized suites data docs93.7%
15Claude 3 Opus· Stanford HELM specialized suites data docs93.3%
16Llama 3.3 70B Instruct Turbo· Stanford HELM specialized suites data docs92.9%
17Nova Prov1· Stanford HELM specialized suites data docs92.4%
18Llama-3.2-90B-Vision-Instruct-Turbo· Stanford HELM specialized suites data docs92%
19Qwen2 72B Instruct· Stanford HELM specialized suites data docs92%
20Meta Llama 3.1 70B Instruct Turbo· Stanford HELM specialized suites data docs91.6%
21Phi-3 Medium 4K Instruct· Stanford HELM specialized suites data docs90.8%
22gemini-1.5-flash-002· Stanford HELM specialized suites data docs90.3%
23Jamba 1.5 Large· Stanford HELM specialized suites data docs90.3%
24gemini-1.5-flash-001· Stanford HELM specialized suites data docs89.9%
25GPT-4o mini· Stanford HELM specialized suites data docs89.9%
26Phi-3 Small 8K Instruct· Stanford HELM specialized suites data docs89.9%
27Mistral Large2407· Stanford HELM specialized suites data docs89.5%
28gemma2-27b· Stanford HELM specialized suites data docs88.7%
29Meta Llama 3 70b· Stanford HELM specialized suites data docs88.7%
30Qwen2.5 7B Instruct Turbo· Stanford HELM specialized suites data docs88.7%
31Qwen1.5-110B-Chat· Stanford HELM specialized suites data docs87%
32mixtral-8x22b· Stanford HELM specialized suites data docs86.6%
33Nova Litev1· Stanford HELM specialized suites data docs86.6%
34Yi-34B· Stanford HELM specialized suites data docs85.7%
35Claude 3 Haiku· Stanford HELM specialized suites data docs85.3%
36Claude 3 Sonnet· Stanford HELM specialized suites data docs85.3%
37Claude 2.1· Stanford HELM specialized suites data docs83.2%
38DBRX-instruct· Stanford HELM specialized suites data docs83.2%
39Qwen1.5-72B· Stanford HELM specialized suites data docs83.2%
40Mistral Largesetting not named by the source· Stanford HELM specialized suites data docs82.8%
41Claude Haiku 3.5· Stanford HELM specialized suites data docs82.4%
42DeepSeek LLM 67B Chat· Stanford HELM specialized suites data docs82.4%
43gemma-2-9b· Stanford HELM specialized suites data docs81.9%
44Command R Plus· Stanford HELM specialized suites data docs80.7%
45Nova Microv1· Stanford HELM specialized suites data docs79.4%
46GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs79%
47Mistral Small· Stanford HELM specialized suites data docs79%
48Claude Instant 1.2· Stanford HELM specialized suites data docs77.3%
49Llama 2 70b· Stanford HELM specialized suites data docs76.9%
50Snowflake Arctic Instruct· Stanford HELM specialized suites data docs76.1%

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