Benchmarks · composite
Arena Text · Japanese
Crowdsourced pairwise human preference on open-ended text prompts (style control on); Bradley-Terry rating. Only the Japanese category.
175 models · 186 results · data as of 2026-10-29 · arena.ai ↗
Leaderboard
Fable 5.1 max
All results
186 rows| # | Model | Effort | Arena Text · Japanese |
|---|---|---|---|
| 1 | Fable 5.1 | max | 1553 ±36 |
| 2 | Opus 5.5 | high | 1535 ±41 |
| 3 | Gemini 4 Argon | high | 1533 ±43 |
| 4 | Fable 5 | high | 1517 ±24 |
| 5 | Gemini 3.7 Flash | high | 1515 ±27 |
| 6 | Kimi K3 | max | 1509 ±24 |
| 7 | Opus 5 | max | 1507 ±26 |
| 8 | Opus 5 | high | 1506 ±20 |
| 9 | GPT-5.5 | high | 1505 ±21 |
| 10 | Gemini 3.8 Flash | high | 1504 ±22 |
| 11 | Gemini 3 Pro | high | 1503 ±30 |
| 12 | Opus 4.6 | high | 1498 ±21 |
| 13 | Gemini 3.1 Pro | high | 1496 ±17 |
| 14 | GPT-5.6 Sol | xhigh | 1495 ±24 |
| 15 | Gemini 3 Flash | high | 1488 ±34 |
| 16 | Opus 4.7 | high | 1488 ±24 |
| 17 | Qwen 3.5 Max | 1484 ±44 | |
| 18 | Gemini 3.5 Flash | high | 1482 ±24 |
| 19 | Muse Spark 1.3 | max | 1482 ±32 |
| 20 | Opus 4.6 | max | 1480 ±20 |
| 21 | GPT-5.5 | xhigh | 1479 ±23 |
| 22 | GPT-5.4 | high | 1478 ±25 |
| 23 | Gemini 3.6 Flash | high | 1477 ±25 |
| 24 | Grok 4.20 | xhigh | 1476 ±40 |
| 25 | Qwen 3.8 Max | max | 1475 ±27 |
| 26 | Muse Spark 1.2 | xhigh | 1473 ±42 |
| 27 | Opus 4.7 | max | 1473 ±23 |
| 28 | GLM 5.3 | max | 1467 ±28 |
| 29 | Muse Spark 1.1 | max | 1465 ±23 |
| 30 | Sonnet 5.5 | xhigh | 1465 ±40 |