| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | InternVL-2-8B | OpenGVLab | 95.9 | 1 | 2026-08-24 |
| 2 | InternVL-2-4B | OpenGVLab | 94.9 | 1 | 2026-08-24 |
| 3 | Phi 3.5 Vision Instruct | Microsoft | 91.3 | 1 | 2026-08-24 |
| 4 | Gemini 1.5 Pro | 86 | 1 | 2026-08-24 | |
| 5 | GPT-4o mini | OpenAI | 84 | 1 | 2026-08-24 |
| 6 | Claude 3.5 Sonnet | Anthropic | 73.8 | 1 | 2026-08-24 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.