LMArena Creative Writing
Text arena rating on creative-writing prompts.
As of 19 Sept 2026, Claude Fable 5 leads LMArena Creative Writing on BenchLeader with 1504, ahead of Claude Opus 4.6 at 1500, across 364 model configurations with a published result.
- Published by
- LMArena
- Category
- Human preference
- Index weight
- Reference only
- Models
- 364
- Data as of
- 19 Sept 2026
CC BY 4.0 — lmarena-ai/leaderboard-dataset on Hugging Face.
What the test looks like
Pairwise votes on stories, poems and other creative prompts.
How it is scored
Bradley-Terry rating from pairwise votes, in Elo-like units, with style control.
What to keep in mind
Taste-driven by design; the style control removes length and formatting effects but not voice preferences.
- 1Claude Fable 51504
- 2Claude Opus 4.6 (high)1500
- 3Gemini 3.8 Flash (high)1496
- 4Gemini 3.7 Flash (high)1495
- 5Claude Opus 4.7 (high)1488
- 6Claude Fable 5.1 (max)1486
- 7Gemini 3 Pro1483
- 8Claude Opus 4.71481
- 9Gemini 3.1 Pro1480
- 10Claude Opus 4.61479
- 11Claude Opus 5 (high)1474
- 12GPT-5.6 Sol (xhigh)1473
- 13Claude Opus 5 (max)1472
- 14Qwen3 8 (max)1472
- 15Gemini 3.6 Flash (high)1470
364 of 364
| # | ||||
|---|---|---|---|---|
| 1 |  Claude Fable 5Anthropic | 1504 | 68.3 | 2026-09-13 |
| 2 | Claude Opus 4.6highAnthropic | 1500 | 61.3 | 2026-09-13 |
| 3 |  Gemini 3.8 FlashhighGoogle | 1496 | 64.5 | 2026-09-13 |
| 4 | Gemini 3.7 FlashhighGoogle | 1495 | 64.6 | 2026-09-13 |
| 5 | Claude Opus 4.7highAnthropic | 1488 | 62.4 | 2026-09-13 |
| 6 | Claude Fable 5.1maxAnthropic | 1486 | 69.7 | 2026-09-13 |
| 7 | Gemini 3 ProGoogle | 1483 | 61.1 | 2026-09-13 |
| 8 | Claude Opus 4.7Anthropic | 1481 | 64.5 | 2026-09-13 |
| 9 | Gemini 3.1 ProGoogle | 1480 | 63.9 | 2026-09-13 |
| 10 | Claude Opus 4.6Anthropic | 1479 | 63.7 | 2026-09-13 |
| 11 | Claude Opus 5highAnthropic | 1474 | 70.2 | 2026-09-13 |
| 12 |  GPT-5.6 SolxhighOpenAI | 1473 | 68.2 | 2026-09-13 |
| 13 | Claude Opus 5maxAnthropic | 1472 | 69.9 | 2026-09-13 |
| 14 |  Qwen3 8maxAlibaba | 1472 | 66.5 | 2026-09-13 |
| 15 | Gemini 3.6 FlashhighGoogle | 1470 | 61.7 | 2026-09-13 |
| 16 | Claude Opus 4.5highAnthropic | 1468 | 57.7 | 2026-09-13 |
| 17 | Gemini 3.5 FlashmediumGoogle | 1466 | 64.2 | 2026-09-13 |
| 18 | Gemini 3.5 FlashhighGoogle | 1465 | 63.6 | 2026-09-13 |
| 19 | Claude Opus 4.8highAnthropic | 1465 | 62.2 | 2026-09-13 |
| 20 |  Muse SparkMeta | 1464 | 65.7 | 2026-09-13 |
| 21 | Claude Opus 4.8Anthropic | 1463 | 61.9 | 2026-09-13 |
| 22 |  GLM-5.3maxZhipu AIopen ↗ | 1462 | 65.8 | 2026-09-13 |
| 23 |  Grok 4.20 Beta1xAI | 1462 | – | 2026-09-13 |
| 24 | GPT-6 AstramaxOpenAI | 1461 | 71.8 | 2026-09-13 |
| 25 | Claude Opus 4.5Anthropic | 1461 | 58.2 | 2026-09-13 |
| 26 | Gemini 3 FlashGoogle | 1459 | 57.7 | 2026-09-13 |
| 27 |  Kimi K3maxMoonshot AIopen ↗ | 1458 | 67.2 | 2026-09-13 |
| 28 | GPT-5.5 InstantOpenAI | 1456 | 59.2 | 2026-09-13 |
| 29 | Claude Sonnet 4.5Anthropic | 1453 | 54.3 | 2026-09-13 |
| 30 | Muse Spark 1.3maxMeta | 1453 | 69.3 | 2026-09-13 |
| 31 | Muse Spark 1.2xhighMeta | 1452 | 64.1 | 2026-09-13 |
| 32 | Grok 4.6highxAI | 1451 | 64.5 | 2026-09-13 |
| 33 | GPT-5.5highOpenAI | 1451 | 67.0 | 2026-09-13 |
| 34 | Muse Spark 1.1Meta | 1451 | 65.2 | 2026-09-13 |
| 35 | GLM-5.2maxZhipu AIopen ↗ | 1451 | 63.9 | 2026-09-13 |
| 36 | Grok 4.5xAI | 1450 | 60.3 | 2026-09-13 |
| 37 | Claude Sonnet 4.6Anthropic | 1450 | 59.0 | 2026-09-13 |
| 38 | Claude Sonnet 4.5highAnthropic | 1449 | 58.8 | 2026-09-13 |
| 39 | GPT-5.5OpenAI | 1448 | 63.2 | 2026-09-13 |
| 40 | Grok 4.20 Multi-AgentxAI | 1448 | 59.2 | 2026-09-13 |
| 41 | GLM-5Zhipu AIopen | 1448 | 54.7 | 2026-09-13 |
| 42 | Qwen3 5maxAlibaba | 1447 | – | 2026-09-13 |
| 43 | GLM-5.1Zhipu AIopen ↗ | 1447 | 56.9 | 2026-09-13 |
| 44 | Grok 4.20thinkingxAI | 1446 | 59.9 | 2026-09-13 |
| 45 | Gemini 3 FlashminimalGoogle | 1446 | 54.9 | 2026-09-13 |
| 46 | Claude Opus 4.1thinkingAnthropic | 1445 | 57.2 | 2026-09-13 |
| 47 | GPT-5.4highOpenAI | 1444 | 59.0 | 2026-09-13 |
| 48 |  DeepSeek V4 ProDeepSeekopen ↗ | 1444 | 52.6 | 2026-09-13 |
| 49 | Qwen3 7maxAlibaba | 1444 | 61.0 | 2026-09-13 |
| 50 | Gemini 2.5 ProGoogle | 1443 | 54.5 | 2026-09-13 |
| 51 | Deepseek v4 ProhighDeepSeek | 1442 | 60.8 | 2026-09-13 |
| 52 | Claude Opus 4.1Anthropic | 1442 | 51.9 | 2026-09-13 |
| 53 | Qwen3 6maxAlibaba | 1438 | 63.3 | 2026-09-13 |
| 54 | Claude Sonnet 5highAnthropic | 1437 | 62.2 | 2026-09-13 |
| 55 | GPT-5.4OpenAI | 1437 | 59.2 | 2026-09-13 |
| 56 | GPT-4.5OpenAI | 1436 | 50.2 | 2026-09-13 |
| 57 | Gemini 3.5 Flash LiteGoogle | 1436 | 55.5 | 2026-09-13 |
| 58 |  MiMo-V2.5-ProXiaomiopen ↗ | 1435 | 59.2 | 2026-09-13 |
| 59 | GLM-5.3-FlashZhipu AIopen ↗ | 1435 | 63.9 | 2026-09-13 |
| 60 | GPT-5.2OpenAI | 1433 | 58.4 | 2026-09-13 |
Cite as: BenchLeader, “LMArena Creative Writing leaderboard”, https://www.benchleader.com/benchmarks/lmarena_text_creative, data as of 19 Sept 2026.
LMArena Creative Writing: questions
- What does LMArena Creative Writing measure?
- Pairwise votes on stories, poems and other creative prompts. Scores are reported in Elo-style ratings from pairwise comparisons; higher is better.
- Which AI model leads LMArena Creative Writing?
- Claude Fable 5 leads LMArena Creative Writing with 1504 as of 19 Sept 2026, ahead of Claude Opus 4.6 at 1500.
- How many models have LMArena Creative Writing results?
- 364 model configurations have a LMArena Creative Writing result on BenchLeader, all taken from LMArena.
- Who runs LMArena Creative Writing and how often is it updated?
- LMArena Creative Writing is published by LMArena. BenchLeader re-reads the published results every morning and records the date each result was published.
- Does LMArena Creative Writing count toward the BenchLeader Index?
- No. LMArena Creative Writing is shown for reference but left out of the composite index.