MEGA-Bench Task - Perception Test Video Action Count: leaderboard
The "Perception Test Video Action Count" 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: Not forecast. 44 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL3-78B | 75 | #352 |
| 2 | InternVL2.5-78B | 68.8 | #431 |
| 3 | Gemini 1.5 Pro (002) | 56.2 | #349 |
| 4 | Gemini 2.5 Pro (Preview 03-25) | 50 | #123 |
| 5 | Qwen 2 VL 7B | 50 | #750 |
| 6 | Aquila-VL-2B | 50 | #1238 |
| 7 | InternVL3-14B | 43.8 | #505 |
| 8 | Qwen 2 VL 72B | 43.8 | #460 |
| 9 | GPT-4o Mini | 37.5 | #599 |
| 10 | InternVL3-38B | 37.5 | #405 |
| 11 | Qwen 2 VL 2B | 37.5 | #1220 |
| 12 | GLM-4.1V-9B (Thinking) | 37.5 | #472 (GLM-4.1V-9B) |
| 13 | Llama 3.2 11B | 37.5 | #1199 |
| 14 | Pixtral-12B | 31.2 | #808 |
| 15 | Gemini 2.0 Flash (Preview) | 31.2 | #370 |
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-perception-test-video-action-count · How It Works · Data refreshed daily, snapshot 2026-10-20.