MLVU-Test - Tutorial QA: leaderboard
MLVU-Test Tutorial QA (TQA): the model must answer questions that need several steps of a long tutorial video (added for the test set), choosing among six options. Scored as accuracy in percent on the held-out long-video test set; higher is better (BAAI, CVPR 2025).
Metric: Multiple-choice accuracy (%). Source: github.com. Saturation forecast: Estimated already saturated. 35 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 48.7 | #741 |
| 2 | Claude 3 Opus | 25.6 | #1459 |
| 3 | Qwen-VL-Max | 22.6 | #1226 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=mlvu-test-tutorial-qa · How It Works · Data refreshed daily, snapshot 2026-11-12.