Open-ended writing on HelloBench
82Average ScoreQwen3-8B + R2-Write-SFT + RLp
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3-8B + R2-Write-SFT + RLpTraining=SFT + PPO2026.04 | 82 | |
| LongWriter-Zero-32BTraining=SFT + GRPO2026.04 | 80.66 | |
| Qwen3-8B + R2-Write-SFTTraining=SFT2026.04 | 80.32 | |
| Qwen3-4B + R2-Write-SFT + RLpTraining=SFT + PPO2026.04 | 79.16 | |
| Qwen3-8B + RLTraining=PPO2026.04 | 76.8 | |
| Qwen3-4B + R2-Write-SFTTraining=SFT2026.04 | 76.36 | |
| Reverse-Engineering-8BTraining=SFT2026.04 | 74.9 | |
| Qwen3-8BTraining=-2026.04 | 71.8 | |
| Qwen3-4B + RLTraining=PPO2026.04 | 70.2 | |
| Qwen3-4BTraining=-2026.04 | 65.86 | |
| Longwriter-9BTraining=SFT2026.04 | 62 |