Benchmarks · composite
Arena Text · Industry: legal and government
Crowdsourced pairwise human preference on open-ended text prompts (style control on); Bradley-Terry rating. Only the Industry: legal and government category.
208 models · 219 results · updated 2026-11-09 · arena.ai ↗
Leaderboard
Gemini 4 Argon high
All results
219 rows| # | Model | Effort | Score |
|---|---|---|---|
| 1 | Gemini 4 Argon | high | 1526 ±16 |
| 2 | Muse Wonder | max | 1524 ±20 |
| 3 | Gemini 4 Carbon | high | 1521 ±37 |
| 4 | Muse Spark 1.2 | xhigh | 1516 ±18 |
| 5 | Opus 4.7 | high | 1512 ±9 |
| 6 | Fable 5 | high | 1511 ±10 |
| 7 | Opus 4.6 | high | 1509 ±8 |
| 8 | GPT-6.1 Sol | max | 1508 ±20 |
| 9 | Opus 4.6 | max | 1508 ±8 |
| 10 | Opus 5.5 | high | 1506 ±19 |
| 11 | Kimi K3 | max | 1506 ±11 |
| 12 | Opus 5 | max | 1502 ±12 |
| 13 | Gemini 3 Pro | high | 1501 ±11 |
| 14 | Muse Spark | 1500 ±19 | |
| 15 | Muse Spark 1.3 | max | 1500 ±14 |
| 16 | Opus 5 | high | 1500 ±9 |
| 17 | Gemini 3.1 Pro | high | 1498 ±7 |
| 18 | Opus 4.8 | max | 1498 ±9 |
| 19 | GPT-5.6 Sol | xhigh | 1497 ±11 |
| 20 | Muse Spark 1.1 | max | 1497 ±10 |
| 21 | Gemini 3.7 Flash | high | 1496 ±12 |
| 22 | GPT-5.5 | high | 1496 ±9 |
| 23 | Fable 5.1 | max | 1496 ±16 |
| 24 | Gemini 3.8 Flash | high | 1496 ±9 |
| 25 | Opus 4.8 | high | 1495 ±9 |
| 26 | Opus 4.7 | max | 1495 ±9 |
| 27 | Kimi K3.1 | max | 1495 ±42 |
| 28 | Qwen 3.8 Max | max | 1495 ±12 |
| 29 | GLM 5.2 | max | 1494 ±10 |
| 30 | GPT-5.4 | high | 1489 ±9 |