Benchmarks · composite
Arena Text · French
Crowdsourced pairwise human preference on open-ended text prompts (style control on); Bradley-Terry rating. Only the French category.
179 models · 190 results · data as of 2026-10-29 · arena.ai ↗
Leaderboard
Gemini 4 Argon high
All results
190 rows| # | Model | Effort | Arena Text · French |
|---|---|---|---|
| 1 | Gemini 4 Argon | high | 1548 ±28 |
| 2 | Fable 5 | high | 1525 ±17 |
| 3 | Muse Spark | 1520 ±28 | |
| 4 | Muse Spark 1.2 | xhigh | 1518 ±31 |
| 5 | Muse Spark 1.3 | max | 1517 ±23 |
| 6 | Opus 4.7 | high | 1517 ±14 |
| 7 | Gemini 3.6 Flash | high | 1513 ±19 |
| 8 | Muse Spark 1.1 | max | 1513 ±16 |
| 9 | Fable 5.1 | max | 1511 ±26 |
| 10 | GLM 5.3 | max | 1510 ±21 |
| 11 | Opus 4.7 | max | 1510 ±14 |
| 12 | Opus 4.6 | high | 1509 ±13 |
| 13 | GPT-6.1 Sol | max | 1507 ±31 |
| 14 | GPT-5.6 Sol | xhigh | 1507 ±17 |
| 15 | Opus 5 | high | 1507 ±14 |
| 16 | Gemini 3 Pro | high | 1506 ±20 |
| 17 | Opus 4.8 | high | 1506 ±14 |
| 18 | Qwen 3.8 Max | max | 1506 ±19 |
| 19 | Opus 4.8 | max | 1506 ±14 |
| 20 | Sonnet 5.5 | xhigh | 1505 ±29 |
| 21 | Kimi K3 | max | 1505 ±17 |
| 22 | Grok 4.20 | xhigh | 1503 ±14 |
| 23 | Gemini 3.1 Pro | high | 1502 ±11 |
| 24 | Gemini 3.7 Flash | high | 1502 ±20 |
| 25 | GPT-5.4 | high | 1501 ±14 |
| 26 | GPT-5.5 | xhigh | 1500 ±14 |
| 27 | Gemini 3.8 Flash | high | 1499 ±16 |
| 28 | GLM 5.3 Flash | max | 1498 ±17 |
| 29 | Opus 4.6 | max | 1495 ±13 |
| 30 | GPT-5.5 | high | 1495 ±14 |