OpenVLM MMT-Bench - Gaze Estimation: leaderboard
The "Gaze Estimation" subtask of OpenVLM MMT-Bench. OpenCompass OpenVLM evaluation of MMT-Bench: multimodal tasks covering visual recognition, localization, OCR, counting, relation reasoning, visual math, and hallucination detection.
Metric: Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 207 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2 VL 2B | 30 | #1220 |
| 2 | Gemma 3 4B | 25 | #1089 |
| 3 | Moonshot v1 8K | 25 | #617 |
| 4 | Llama 3.2 90B Vision Instruct | 25 | #622 |
| 5 | InternVL3-78B | 20 | #352 |
| 6 | GPT-4.1 (2025-04-14) | 20 | #250 |
| 7 | InternVL3-14B | 20 | #505 |
| 8 | Gemini 1.5 Pro | 20 | #445 |
| 9 | Qwen 2.5 VL 7B | 20 | #643 |
| 10 | Llama 3.2 11B Instruct | 20 | #1126 |
| 11 | Claude 3 Haiku | 20 | #803 |
| 12 | GPT-4.1 Nano | 15 | #731 |
| 13 | InternVL3-8B | 15 | #620 |
| 14 | Gemma 3 27B | 15 | #607 |
| 15 | Qwen 2 VL 72B | 15 | #460 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmt-bench-gaze-estimation · How It Works · Data refreshed daily, snapshot 2026-10-20.