Writing Evaluation on HelloBench-G
62.89ScoreGemini-2.5-pro
Evaluation results
| Method | Links | |
|---|---|---|
| Gemini-2.5-pro2026.06 | 62.89 | |
| MiniMax-M2.52026.06 | 59.33 | |
| OPERABackbone=Llama3.1-8B2026.06 | 49.83 | |
| OPERABackbone=Qwen3-8B2026.06 | 49.79 | |
| Qwen3-8B-SFTMode=SFT2026.06 | 41.96 | |
| GPT-4o-05132026.06 | 38.02 | |
| Llama3.1-8B-SFTMode=SFT2026.06 | 37.53 | |
| Qwen3-8B2026.06 | 37.12 | |
| DeepWriter-8B2026.06 | 0.25 | |
| Llama3.1-8B2026.06 | -26.07 | |
| LongWriter-Zero-32B2026.06 | -55.48 | |
| LongWriter-8B2026.06 | -88.45 |