Chatbot Arena (Text - Instruction Following): leaderboard
The arena.ai text board restricted to prompts that give clear actionable instructions, often with formatting or structural requirements, so the battle turns on carrying them out precisely rather than on domain knowledge. A language model classifier applies the tag and voters pick between two anonymous replies. The rating is refit over this subset alone, and higher is better.
Metric: Arena Score. Source: arena.ai. 414 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon (High) | 1529 | #8 (Gemini 4 Argon) |
| 2 | Claude Opus 5.5 (High) | 1519 | #3 (Claude Opus 5.5) |
| 3 | Claude Opus 4.6 (High) | 1513 | #63 (Claude Opus 4.6) |
| 4 | Claude Fable 5 (High) | 1509 | #16 (Claude Fable 5) |
| 5 | Claude Opus 4.7 (High) | 1502 | #49 (Claude Opus 4.7) |
| 6 | Claude Opus 4.6 | 1499 | #63 |
| 7 | Claude Fable 5.1 (Max) | 1498 | #9 (Claude Fable 5.1) |
| 8 | Claude Opus 5 (High) | 1497 | #12 (Claude Opus 5) |
| 9 | Claude Opus 4.7 | 1493 | #49 |
| 10 | Claude Opus 5 (Max) | 1493 | #12 (Claude Opus 5) |
| 11 | GPT-5.6 Sol (xHigh) | 1490 | #19 (GPT-5.6 Sol) |
| 12 | Claude Opus 4.8 (High) | 1490 | #37 (Claude Opus 4.8) |
| 13 | Claude Sonnet 5.5 (xHigh) | 1489 | #11 (Claude Sonnet 5.5) |
| 14 | Kimi K3 (Max) | 1488 | #30 (Kimi K3) |
| 15 | Gemini 3.8 Flash (High) | 1488 | #21 (Gemini 3.8 Flash) |
Interactive version: theaggregate.ai/benchmark?slug=chatbot-arena-text-instruction-following · How It Works · Data refreshed daily, snapshot 2026-10-20.