| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Gemini 1.5 Pro | 89.3 | 1 | 2026-08-24 | |
| 2 | Phi 3.5 Vision Instruct | Microsoft | 86.1 | 1 | 2026-08-24 |
| 3 | InternVL-2-8B | OpenGVLab | 84.2 | 1 | 2026-08-24 |
| 4 | GPT-4o mini | OpenAI | 83.6 | 1 | 2026-08-24 |
| 5 | InternVL-2-4B | OpenGVLab | 83.3 | 1 | 2026-08-24 |
| 6 | Claude 3.5 Sonnet | Anthropic | 76.6 | 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.