MEGA-Bench Task - AV Multicamera Tracking Predict BBox: leaderboard

The "AV Multicamera Tracking Predict BBox" 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

#ModelScoreOverall rank
1Gemini 2.5 Pro (Preview 03-25)13.4#123
2Qwen 2 VL 72B8#460
3Gemma 3 27B (IT)5.4#523
4GPT-4o Mini4.7#599
5Gemini 2.0 Flash (Preview)4.5#370
6GPT-4o3.8#343
7Claude 3.5 Sonnet (20240620)3.8#299
8MiniCPM-V-2.63.4#836
9Qwen 2 VL 2B3.4#1220
10Aquila-VL-2B3.4#1238
11InternVL2-8B3.2#817
12Llama 3.2 11B3.2#1199
13Qwen 2 VL 7B2.7#750
14Gemma 3 12B (IT)2.4#664
15Claude 3.5 Sonnet (20241022)2.3#291

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-av-multicamera-tracking-predict-bbox · How It Works · Data refreshed daily, snapshot 2026-10-20.