Human-AI Co-writing on CoAuthor (Overall)
19.61HAR (%)GPT-5.1
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-5.1Level=L22026.05 | 19.61 | 10.5 | |
| Gemini-2.5-ProLevel=L22026.05 | 17.89 | 9.39 | |
| GPT-5.1Level=L12026.05 | 17.59 | 11.28 | |
| Gemini-2.5-FlashLevel=L22026.05 | 16.05 | 10.24 | |
| Gemini-2.5-ProLevel=L12026.05 | 14.17 | 9.17 | |
| Gemini-2.5-FlashLevel=L12026.05 | 13.39 | 9.68 | |
| DeepSeek-V3.2Level=L22026.05 | 10.37 | 9.97 | |
| DeepSeek-V3.2Level=L12026.05 | 6.99 | 10.18 |