Benchmarks · composite
Arena Text · Polish
Crowdsourced pairwise human preference on open-ended text prompts (style control on); Bradley-Terry rating. Only the Polish category.
155 models · 166 results · updated 2026-11-09 · arena.ai ↗
Leaderboard
Muse Wonder max
All results
166 rows| # | Model | Effort | Arena Text · Polish |
|---|---|---|---|
| 1 | Muse Wonder | max | 1547 ±42 |
| 2 | Gemini 4 Argon | high | 1527 ±36 |
| 3 | Fable 5.1 | max | 1527 ±33 |
| 4 | Muse Spark 1.2 | xhigh | 1523 ±40 |
| 5 | Muse Spark 1.1 | max | 1513 ±22 |
| 6 | Fable 5 | high | 1510 ±22 |
| 7 | Opus 5.5 | high | 1505 ±40 |
| 8 | Gemini 3 Pro | high | 1505 ±17 |
| 9 | Gemini 3.7 Flash | high | 1504 ±26 |
| 10 | Gemini 3.8 Flash | high | 1504 ±20 |
| 11 | Gemini 3.5 Flash | high | 1503 ±21 |
| 12 | Opus 4.7 | max | 1502 ±18 |
| 13 | Opus 4.6 | high | 1502 ±15 |
| 14 | Opus 5 | max | 1501 ±25 |
| 15 | Opus 5 | high | 1501 ±18 |
| 16 | GPT-5.6 Sol | xhigh | 1500 ±23 |
| 17 | Muse Spark 1.3 | max | 1498 ±31 |
| 18 | Gemini 3 Flash | high | 1497 ±19 |
| 19 | Gemini 3.5 Flash | medium | 1496 ±23 |
| 20 | Gemini 3.1 Pro | high | 1496 ±12 |
| 21 | Sonnet 5.5 | xhigh | 1495 ±38 |
| 22 | Opus 4.6 | max | 1494 ±15 |
| 23 | Opus 4.7 | high | 1490 ±17 |
| 24 | Opus 4.8 | high | 1487 ±18 |
| 25 | Muse Spark | 1487 ±33 | |
| 26 | Grok 4.1 | 1486 ±14 | |
| 27 | Kimi K3 | max | 1486 ±22 |
| 28 | GPT-6.1 Sol | max | 1485 ±42 |
| 29 | Opus 4.8 | max | 1484 ±17 |
| 30 | GPT-5.4 | high | 1483 ±17 |