Skip to content

CC-OCR

An OCR-focused benchmark for reading and extracting text from visually complex documents and images.

7 models

Higher is better.

  1. 1Qwen3.5 397BAlibaba82.0
  2. 2Qwen3.6-35B-A3BAlibaba81.9
  3. 3Qwen3.6-27BAlibaba81.2
  4. 4Kimi K2.5Moonshot AI79.7
  5. 5Qwen3.8 MaxAlibaba79.6
  6. 6Gemini 3 ProGoogle79.0
  7. 7GPT-5.2OpenAI70.3
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Text extraction from images and documents
Tasks
Optical character recognition
Authors
Qwen (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.