Benchmarks · composite
Arena Text · Industry: writing, literature and language
Crowdsourced pairwise human preference on open-ended text prompts (style control on); Bradley-Terry rating. Only the Industry: writing, literature and language category.
212 models · 223 results · updated 2026-11-09 · arena.ai ↗
Leaderboard
Gemini 4 Carbon high
All results
223 rows| # | Model | Effort | Score |
|---|---|---|---|
| 1 | Gemini 4 Carbon | high | 1529 ±18 |
| 2 | Gemini 4 Argon | high | 1526 ±9 |
| 3 | Opus 5.5 | high | 1517 ±10 |
| 4 | Fable 5 | high | 1509 ±6 |
| 5 | Opus 4.6 | high | 1500 ±5 |
| 6 | Muse Wonder | max | 1499 ±11 |
| 7 | Fable 5.1 | max | 1497 ±9 |
| 8 | Opus 4.7 | high | 1497 ±6 |
| 9 | Gemini 3.8 Flash | high | 1493 ±6 |
| 10 | Gemini 3.7 Flash | high | 1490 ±7 |
| 11 | Opus 4.6 | max | 1489 ±5 |
| 12 | Opus 4.7 | max | 1483 ±6 |
| 13 | GPT-5.6 Sol | xhigh | 1483 ±7 |
| 14 | Gemini 3.1 Pro | high | 1482 ±5 |
| 15 | Gemini 3 Pro | high | 1481 ±7 |
| 16 | Opus 5 | high | 1481 ±6 |
| 17 | Opus 5 | max | 1481 ±7 |
| 18 | Muse Spark 1.3 | max | 1478 ±8 |
| 19 | Opus 4.8 | high | 1476 ±6 |
| 20 | Gemini 3.5 Flash | high | 1475 ±6 |
| 21 | Qwen 3.8 Max | max | 1475 ±7 |
| 22 | Gemini 3.6 Flash | high | 1474 ±7 |
| 23 | Kimi K3 | max | 1473 ±7 |
| 24 | Gemini 3.5 Flash | medium | 1472 ±6 |
| 25 | Kimi K3.1 | max | 1470 ±19 |
| 26 | GPT-6.1 Sol | max | 1469 ±11 |
| 27 | GPT-5.5 | high | 1468 ±6 |
| 28 | Sonnet 5.5 | xhigh | 1468 ±10 |
| 29 | Opus 4.8 | max | 1468 ±6 |
| 30 | GLM 5.3 | max | 1468 ±7 |