OpenVLM SEED-Bench 2: leaderboard
OpenCompass OpenVLM evaluation of SEED-Bench 2: a hierarchical multimodal benchmark testing image, video, and interleaved image-text understanding across 27 dimensions.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 59 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 (2025-04-14) | 76 | #250 |
| 2 | GPT-4.1 Mini | 75 | #355 |
| 3 | Gemini 1.5 Flash | 70.2 | #618 |
| 4 | InternVL3-14B | 67.7 | #505 |
| 5 | InternVL3-8B | 67.1 | #620 |
| 6 | GPT-4.1 Nano | 64.4 | #731 |
| 7 | Gemma 3 27B | 59.9 | #607 |
| 8 | Pixtral-12B | 57.9 | #808 |
| 9 | Gemma 3 12B | 57.6 | #671 |
| 10 | Gemma 3 4B | 52.7 | #1089 |
| 11 | Aquila-VL-2B | 49.3 | #1238 |
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 · How It Works · Data refreshed daily, snapshot 2026-10-20.