Roboflow Vision Evals — OCR Single Value (October)
Multimodal
Exact-match score for single readings in the October OCR suite. Single preliminary run per effort tier; question count unreported.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 3.7 Flash | 74.8 |
| Gemini 3.8 Flash | 73.5 |
| Claude Fable 5 | 72.2 |
| Gemini 3.5 Flash | 71.7 |
| Claude Opus 5.5 | 69.6 |
| Claude Fable 5.1 | 69.1 |
| Gemini 3 Flash Preview | 69.1 |
| Gemini 3.6 Flash | 68.7 |
| Gemini 3.1 Pro Preview | 68.7 |
| Gemini 3.5 Flash-Lite | 66.1 |
| Gemini 2.5 Pro | 63.0 |
| Muse Spark 1.2 | 61.7 |
| GPT-6 Astra | 60.9 |
| Muse Spark 1.1 | 60.4 |
| GPT-6.1 Sol | 58.3 |
| Qwen3.7-Plus (source unspecified) | 58.3 |
| Claude Opus 5 | 56.1 |
| Qwen3.7 Flash | 56.1 |
| Opus 4.8 | 54.8 |
| Sonnet 5 | 54.8 |
| Grok 4.7 | 54.4 |
| Muse Glimmer 30B | 54.4 |
| Qwen3.8 Flash | 51.3 |
| Grok 4.6 | 50.9 |
| GLM-5.3 Flash | 50.9 |
| Qwen3.5-27B | 50.4 |
| Claude Sonnet 5.5 | 48.7 |
| Qwen3-VL-235B-A22B-Instruct | 48.3 |
| Grok 4.5 | 47.8 |
| GPT-6 Sol | 47.4 |
| GPT-5.4 Mini | 47.0 |
| Claude Haiku 5.5 | 46.5 |
| GPT-5.6 Sol | 46.1 |
| GPT-6 Luna | 44.4 |
| GLM 5V Turbo | 44.4 |
| Kimi K3 | 43.9 |
| GPT-5.5 | 43.9 |
| MiMo-V2.6-Flash | 42.2 |
| MiMo-V2.6-Pro | 40.4 |
| Kimi K2.6 | 38.7 |