Human-AI Co-writing on CoAuthor Scientific
23.62HAR (%)GPT-5.1
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-5.1Level=L22026.05 | 23.62 | 12.14 | |
| GPT-5.1Level=L12026.05 | 21.47 | 12.05 | |
| Gemini-2.5-FlashLevel=L22026.05 | 20.04 | 11.79 | |
| Gemini-2.5-ProLevel=L22026.05 | 19.12 | 10.81 | |
| Gemini-2.5-ProLevel=L12026.05 | 15.15 | 11 | |
| Gemini-2.5-FlashLevel=L12026.05 | 14.83 | 11.11 | |
| DeepSeek-V3.2Level=L22026.05 | 12.88 | 11.39 | |
| DeepSeek-V3.2Level=L12026.05 | 8.18 | 11.7 |