OpenVLM SEED-Bench 2 Plus: leaderboard
OpenCompass OpenVLM evaluation of SEED-Bench-2-Plus: image and video understanding across charts, maps, webpages, navigation, OCR, spatial reasoning, and multimodal knowledge.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 211 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 (2025-04-14) | 73.1 | #250 |
| 2 | Qwen 2 VL 72B | 72.3 | #460 |
| 3 | GPT-4.1 Mini | 71.9 | #355 |
| 4 | InternVL3-38B | 71.8 | #405 |
| 5 | InternVL3-78B | 71.7 | #352 |
| 6 | Claude 3.5 Sonnet | 71.7 | #350 |
| 7 | Gemini 1.5 Pro | 70.8 | #445 |
| 8 | Qwen 2.5 VL 7B | 70.7 | #643 |
| 9 | InternVL3-14B | 70.1 | #505 |
| 10 | InternVL3-8B | 69.4 | #620 |
| 11 | Qwen 2 VL 7B | 68.6 | #750 |
| 12 | Gemini 1.5 Flash | 68.6 | #618 |
| 13 | Llama 3.2 90B Vision Instruct | 68.2 | #622 |
| 14 | Claude 3.7 Sonnet | 67.6 | #249 |
| 15 | Pixtral-12B | 67.4 | #808 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=openvlm-seed-bench-2-plus · How It Works · Data refreshed daily, snapshot 2026-10-20.