Skip to content

Benchmark

HELM MMLU: Marketing - EM

Scores are percentages. Higher is better.

Models measured
64
Scores recorded
68
Sources
1
IQ weight
2.09

Score distribution

90%
Mean
91.9%
Median
8.9%
Std dev
84.2%
10th percentile
94.9%
90th percentile
26.9%
Min
26.9%96.2%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.09 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
1gemini-1.5-pro-002· Stanford HELM specialized suites data docs96.2%
2GPT 4· Stanford HELM specialized suites data docs96.2%
3Llama 3.1 405B Instruct Turbo· Stanford HELM specialized suites data docs96.2%
4Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs95.3%
5gemini-1.5-flash-002· Stanford HELM specialized suites data docs95.3%
6Qwen2 72B Instruct· Stanford HELM specialized suites data docs95.3%
7Qwen2.5 72B Instruct Turbo· Stanford HELM specialized suites data docs95.3%
8Claude 3.5 Sonnetsetting not named by the source· Stanford HELM specialized suites data docs94.9%
9DeepSeek V3· Stanford HELM specialized suites data docs94.9%
10GPT-4 Turbo· Stanford HELM specialized suites data docs94.9%
11Qwen1.5-72B· Stanford HELM specialized suites data docs94.9%
12Claude 3 Opus· Stanford HELM specialized suites data docs94.4%
13Gemini 2.0 Flash Experimental· Stanford HELM specialized suites data docs94.4%
14Llama-3.2-90B-Vision-Instruct-Turbo· Stanford HELM specialized suites data docs94.4%
15DBRX-instruct· Stanford HELM specialized suites data docs94%
16gemini-1.5-flash-001· Stanford HELM specialized suites data docs94%
17gemma2-27b· Stanford HELM specialized suites data docs94%
18GPT 4osetting not named by the source· Stanford HELM specialized suites data docs94%
19Jamba 1.5 Large· Stanford HELM specialized suites data docs94%
20Meta Llama 3 70b· Stanford HELM specialized suites data docs94%
21Mistral Large2407· Stanford HELM specialized suites data docs94%
22text-unicorn· Stanford HELM specialized suites data docs94%
23GPT 4osetting not named by the source· Stanford HELM specialized suites data docs93.6%
24Meta Llama 3.1 70B Instruct Turbo· Stanford HELM specialized suites data docs93.6%
25Yi-34B· Stanford HELM specialized suites data docs93.6%
26gemini-1.5-pro-001· Stanford HELM specialized suites data docs93.2%
27GPT 4 1106 Preview· Stanford HELM specialized suites data docs93.2%
28Command R Plus· Stanford HELM specialized suites data docs92.7%
29GPT-4o mini· Stanford HELM specialized suites data docs92.7%
30Llama 3.3 70B Instruct Turbo· Stanford HELM specialized suites data docs92.7%
31Claude 2.1· Stanford HELM specialized suites data docs92.3%
32DeepSeek LLM 67B Chat· Stanford HELM specialized suites data docs92.3%
33Nova Prov1· Stanford HELM specialized suites data docs92.3%
34gemma-2-9b· Stanford HELM specialized suites data docs91.9%
35Phi-3 Medium 4K Instruct· Stanford HELM specialized suites data docs91.9%
36Qwen1.5-110B-Chat· Stanford HELM specialized suites data docs91.9%
37Qwen2.5 7B Instruct Turbo· Stanford HELM specialized suites data docs91.9%
38Jamba 1.5 Mini· Stanford HELM specialized suites data docs91.5%
39mixtral-8x22b· Stanford HELM specialized suites data docs91.5%
40Claude 3 Haiku· Stanford HELM specialized suites data docs91%
41GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs91%
42GPT-3.5 Turbosetting not named by the source· Stanford HELM specialized suites data docs91%
43Nova Microv1· Stanford HELM specialized suites data docs91%
44Mistral Small· Stanford HELM specialized suites data docs90.6%
45Snowflake Arctic Instruct· Stanford HELM specialized suites data docs90.2%
46Claude Haiku 3.5· Stanford HELM specialized suites data docs89.7%
47Mistral Largesetting not named by the source· Stanford HELM specialized suites data docs89.7%
48Phi-3 Small 8K Instruct· Stanford HELM specialized suites data docs89.7%
49Qwen1.5-14B· Stanford HELM specialized suites data docs89.3%
50Yi-6B· Stanford HELM specialized suites data docs89.3%

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