HELM VHELM - Mmmu - Marketing: leaderboard
VHELM (Stanford CRFM HELM for vision-language models, run v2.1.2) MMMU Marketing: college-level multiple-choice questions over images from the MMMU validation set for this subject; 29 questions, prefix quasi-exact match. Percent correct, 0-100, higher is better; a small sample.
Metric: Prefix Quasi-Exact Match (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 48 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 96.55 | #126 |
| 2 | O4 Mini (2025-04-16) | 96.55 | #183 |
| 3 | O3 (2025-04-16) | 96.55 | #125 |
| 4 | O1 (2024-12-17) | 89.66 | #157 |
| 5 | Llama 4 Maverick Instruct FP8 | 82.76 | #424 |
| 6 | Gemini 1.5 Pro (002) | 79.31 | #368 |
| 7 | Claude 3.5 Sonnet (20240620) | 68.97 | #308 |
| 8 | Gemini 2.0 Flash (Preview) | 68.97 | #376 |
| 9 | Claude 3.5 Sonnet (20241022) | 65.52 | #302 |
| 10 | GPT-4.5 | 65.52 | #278 |
| 11 | GPT-4.1 Mini | 62.07 | #363 |
| 12 | GPT-4.1 (2025-04-14) | 62.07 | #257 |
| 13 | GPT-4o (2024-08-06) | 62.07 | #343 |
| 14 | Gemini 2.0 Pro (Preview 02-05) | 62.07 | #318 |
| 15 | Gemini 2.0 Flash Lite | 62.07 | #455 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-vhelm-mmmu-marketing · How It Works · Data refreshed daily, snapshot 2026-11-05.