Chatbot Arena (Vision - OCR): leaderboard
The arena.ai vision board restricted to images that contain text, where the model has to read that text and use it to answer the prompt. Voters compare two anonymous answers to the same image and prompt. The rating is refit over the conversations in this category alone, and higher is better.
Metric: Arena Score. Source: arena.ai. 121 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 (High) | 1326 | #16 (Claude Fable 5) |
| 2 | Qwen 3.8 Max | 1317 | #35 |
| 3 | Gemini 3.8 Flash (High) | 1317 | #21 (Gemini 3.8 Flash) |
| 4 | Claude Opus 4.7 | 1313 | #49 |
| 5 | Claude Opus 4.6 (High) | 1313 | #63 (Claude Opus 4.6) |
| 6 | Claude Opus 4.6 | 1312 | #63 |
| 7 | Gemini 3.7 Flash (High) | 1311 | #23 (Gemini 3.7 Flash) |
| 8 | Claude Opus 4.7 (High) | 1311 | #49 (Claude Opus 4.7) |
| 9 | Muse Spark 1.2 (xHigh) | 1307 | #62 (Muse Spark 1.2) |
| 10 | Claude Opus 5.5 (High) | 1307 | #3 (Claude Opus 5.5) |
| 11 | Muse Spark 1.3 (Max) | 1306 | #24 (Muse Spark 1.3) |
| 12 | Claude Opus 5 (High) | 1304 | #12 (Claude Opus 5) |
| 13 | Gemini 3 Pro | 1302 | #80 |
| 14 | GPT-6 Astra (Max) | 1302 | #6 (GPT-6 Astra) |
| 15 | Claude Fable 5.1 (Max) | 1302 | #9 (Claude Fable 5.1) |
Interactive version: theaggregate.ai/benchmark?slug=chatbot-arena-vision-ocr · How It Works · Data refreshed daily, snapshot 2026-10-20.