Instruction Following
LiveBench · 28 models
| Rank | Model | Score |
|---|---|---|
| 1 | gemini-3.1-pro-preview-high | 79.1 |
| 2 | gemini-3.5-flash-high | 75.6 |
| 3 | qwen3.7-max | 74.0415 |
| 4 | claude-opus-4-8-xhigh-effort | 72.371 |
| 5 | claude-fable-5-xhigh-effort | 71.9958 |
| 6 | gpt-5.5-high | 71.375 |
| 7 | gpt-5.5-xhigh | 70.7293 |
| 8 | gpt-5.4-xhigh | 70.2167 |
| 9 | gpt-5.4-nano-xhigh | 67.2048 |
| 10 | claude-opus-4-7-xhigh-effort | 66.7375 |
| 11 | gpt-5.2-codex | 66.4457 |
| 12 | gpt-5.5 | 65.7375 |
| 13 | grok-build-0.1 | 65.2208 |
| 14 | kimi-k2.6-thinking | 64.3583 |
| 15 | claude-sonnet-5-xhigh-effort | 63.8585 |
| 16 | claude-opus-4-6-thinking-auto-high-effort | 63.3125 |
| 17 | claude-sonnet-4-6-thinking-auto-medium-effort | 63.2208 |
| 18 | deepseek-v4-flash | 63.1375 |
| 19 | grok-4.3 | 62.75 |
| 20 | claude-opus-4-5-20251101-thinking-64k-high-effort | 62.546 |
| 21 | deepseek-v4-pro | 62.35 |
| 22 | glm-5.2 | 62.2918 |
| 23 | gpt-5.2-2025-12-11-high | 61.7707 |
| 24 | gpt-5.4-mini-xhigh | 59.804 |
| 25 | qwen3.6-plus | 58.342 |
| 26 | minimax-m3 | 57.5083 |
| 27 | kimi-k2.7-code | 56.2918 |
| 28 | qwen3.6-27b | 53.229 |