Long-form writing on WritingBench D4 English literature/arts
9.5Score (D4 Lit/Arts)GPT-5.4
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5.42026.06 | 9.5 | |
| Claude Opus 4.62026.06 | 9.3 | |
| Gemini 3.1 Pro2026.06 | 9.2 | |
| Gemini 3 Flash2026.06 | 9.1 | |
| Gemma-4-31B2026.06 | 8.7 | |
| Gemma-4-26B-A4BMoE=true2026.06 | 8.5 | |
| Qwen3.5-27B2026.06 | 8.1 | |
| POLARISBackbone=Qwen3.5-9B2026.06 | 7.9 | |
| Gemma-4-E4BPLE=true2026.06 | 7.7 | |
| Qwen3.5-9B2026.06 | 6.8 | |
| Ministral-3-8B2026.06 | 4.8 | |
| LongWriter-Zero-32B2026.06 | 4.6 | |
| DeepSeek-R1-Distill-Qwen-14B2026.06 | 3.9 | |
| LongWriter-Llama3.1-8B2026.06 | 3.9 |