Code Completion
LiveBench · 28 models
| Rank | Model | Score |
|---|---|---|
| 1 | gpt-5.2-codex | 86.957 |
| 2 | claude-fable-5-xhigh-effort | 84.783 |
| 3 | gpt-5.5 | 82.609 |
| 3 | gpt-5.5-high | 82.609 |
| 3 | gpt-5.5-xhigh | 82.609 |
| 6 | claude-opus-4-5-20251101-thinking-64k-high-effort | 80.435 |
| 6 | glm-5.2 | 80.435 |
| 6 | gpt-5.4-xhigh | 80.435 |
| 9 | claude-opus-4-7-xhigh-effort | 78.261 |
| 9 | claude-opus-4-8-xhigh-effort | 78.261 |
| 9 | claude-sonnet-4-6-thinking-auto-medium-effort | 78.261 |
| 9 | claude-sonnet-5-xhigh-effort | 78.261 |
| 9 | gemini-3.1-pro-preview-high | 78.261 |
| 9 | kimi-k2.6-thinking | 78.261 |
| 15 | claude-opus-4-6-thinking-auto-high-effort | 76.087 |
| 15 | gemini-3.5-flash-high | 76.087 |
| 15 | gpt-5.2-2025-12-11-high | 76.087 |
| 15 | kimi-k2.7-code | 76.087 |
| 15 | qwen3.6-plus | 76.087 |
| 20 | gpt-5.4-mini-xhigh | 75.556 |
| 21 | qwen3.6-27b | 71.739 |
| 22 | deepseek-v4-pro | 69.565 |
| 22 | qwen3.7-max | 69.565 |
| 24 | gpt-5.4-nano-xhigh | 67.391 |
| 24 | grok-build-0.1 | 67.391 |
| 24 | minimax-m3 | 67.391 |
| 27 | deepseek-v4-flash | 65.217 |
| 27 | grok-4.3 | 65.217 |