LiveBench Coding Completion — leaderboard
Metric: Score. Source: livebench.ai. 65 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude 3.5 Sonnet (20240620) | 68.42 |
| 2 | GPT-4o (2024-08-06) | 55.26 |
| 3 | GPT-4o ChatGPT | 50 |
| 4 | DeepSeek V2.5 | 50 |
| 5 | Gemini 1.5 Pro (Preview 0827) | 50 |
| 6 | Llama 3.1 405B Instruct | 47.37 |
| 7 | GPT-4o Mini (2024-07-18) | 44.74 |
| 8 | GPT-4o (2024-05-13) | 44.74 |
| 9 | Mistral Large 2 (Jul) | 44.74 |
| 10 | Gemini 1.5 Pro (Preview 0801) | 44.74 |
| 11 | GPT-4 Turbo | 42.1 |
| 12 | Claude 3 Opus (20240229) | 42.1 |
| 13 | GPT-4 (0613) | 42.1 |
| 14 | GPT-4 Preview (0125) | 42.1 |
| 15 | DeepSeek Coder V2 | 42.1 |
Interactive version: aibenchmarks.dev/benchmark?slug=livebench-coding-completion · How the rankings work · Data refreshed daily, snapshot 2026-07-09.