MEGA-Bench Task - Video Detail Description: leaderboard
The "Video Detail Description" subtask of MEGA-Bench Task. A single task from MEGA-Bench, a large-scale multimodal evaluation suite that scores vision-language models across 500+ real-world tasks spanning perception, reasoning, OCR, charts, documents, and multi-image understanding.
Metric: Task Score (%). Source: huggingface.co. Saturation forecast: Around January 2027. 44 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-4.1V-9B (Thinking) | 65.8 | #472 (GLM-4.1V-9B) |
| 2 | Gemini 1.5 Pro (002) | 59.5 | #349 |
| 3 | Gemini 2.5 Pro (Preview 03-25) | 57.9 | #123 |
| 4 | GPT-4o | 56.8 | #343 |
| 5 | InternVL3-8B | 56.3 | #620 |
| 6 | Gemini 1.5 Flash (002) | 56.3 | #583 |
| 7 | InternVL3-14B | 53.7 | #505 |
| 8 | Gemma 3 27B (IT) | 53.2 | #523 |
| 9 | Gemini 2.0 Flash (Preview) | 52.1 | #370 |
| 10 | Gemma 3 12B (IT) | 51.6 | #664 |
| 11 | InternVL3-38B | 50.5 | #405 |
| 12 | InternVL3-78B | 49.5 | #352 |
| 13 | Qwen 2 VL 72B | 47.9 | #460 |
| 14 | Llama 4 Scout Base | 46.8 | #707 |
| 15 | Pixtral-12B | 45.3 | #808 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-video-detail-description · How It Works · Data refreshed daily, snapshot 2026-10-20.