| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | InternVL3-2B | OpenGVLab | 81.1 | 1 | 2026-08-09 |
| 2 | Qwen2.5 VL 3B | Alibaba | 80.4 | 1 | 2026-08-09 |
| 3 | LFM2-VL-3B | Liquid AI | 79.8 | 1 | 2026-08-09 |
| 4 | InternVL3_5-2B | OpenGVLab | 78.2 | 1 | 2026-08-09 |
| 5 | LFM2.5-VL-450M-Extract | Liquid AI | 60.9 | 1 | 2026-08-09 |
| 6 | LFM2-VL-450M | Liquid AI | 56.3 | 1 | 2026-08-09 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.