Vector Eval - MMMU-OE: leaderboard
Vector Institute Inspect Evals run of MMMU open-ended tasks, evaluating multimodal academic reasoning when models must produce free-form answers.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O1 | 69.81 | #183 |
| 2 | Claude 3.5 Sonnet | 41.51 | #350 |
| 3 | GPT-4o | 35.85 | #343 |
| 4 | Gemini 1.5 Pro | 35.85 | #445 |
| 5 | GPT-4o Mini | 18.87 | #599 |
| 6 | Gemini 1.5 Flash | 16.98 | #618 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=vector-eval-mmmu-oe · How It Works · Data refreshed daily, snapshot 2026-10-20.