New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
WritingBench benchmark
AI model leaderboard for the WritingBench benchmark. Compare how large language models score on WritingBench, see the full ranking, and understand what this AI benchmark measures. Qwen3-Next-80B-A3B currently leads with 85.2. Long-form writing benchmark — open-ended creative tasks judged by LLM rubric.
Leaderboard
| # | Model | Organization | Score | Variant | Source |
|---|---|---|---|---|---|
| #1 | Qwen3-Next-80B-A3B | Qwen | 85.2 | from-hf-readme | official ↗ |
| #2 | Qwen3-235B-A22B | Qwen | 8.03 | thinking | official ↗ |
| #3 | Qwen3-32B | Qwen | 7.9 | thinking | official ↗ |
| #4 | Qwen3-14B | Qwen | 7.8 | thinking | official ↗ |
| #5 | Qwen3-235B-A22B | Qwen | 7.7 | non-thinking | official ↗ |
| #6 | Qwen3-30B-A3B | Qwen | 7.7 | thinking | official ↗ |
| #7 | Qwen3-8B | Qwen | 7.59 | thinking | official ↗ |
| #8 | Qwen3-32B | Qwen | 7.54 | non-thinking | official ↗ |
| #9 | Qwen3-4B | Qwen | 7.35 | thinking | official ↗ |
| #10 | Qwen3-14B | Qwen | 7.24 | non-thinking | official ↗ |
| #11 | Qwen3-30B-A3B | Qwen | 7.22 | non-thinking | official ↗ |
| #12 | Qwen3-8B | Qwen | 7.15 | non-thinking | official ↗ |
| #13 | Qwen3-1.7B | Qwen | 7.02 | thinking | official ↗ |
| #14 | Qwen3-4B | Qwen | 6.85 | non-thinking | official ↗ |
| #15 | Qwen3-1.7B | Qwen | 6.54 | non-thinking | official ↗ |
| #16 | Qwen3-0.6B | Qwen | 5.61 | thinking | official ↗ |
| #17 | Qwen3-0.6B | Qwen | 5.13 | non-thinking | official ↗ |
Frequently asked questions about WritingBench
What is the WritingBench benchmark?
Long-form writing benchmark — open-ended creative tasks judged by LLM rubric.
How is the WritingBench benchmark scored?
WritingBench is scored using the rubric_score metric, where a higher score is better. GenAIList aggregates reported scores from model providers and papers into a single ranked leaderboard.
Which AI model scores highest on WritingBench?
As of the latest reported scores on GenAIList, Qwen3-Next-80B-A3B achieves the highest result on WritingBench with a score of 85.2.
Is a higher WritingBench score better?
Yes. On WritingBench a higher score indicates better performance, so models near the top of the leaderboard are the strongest.