LiveBench · Coding
The Frontier
Best score over time · one chart, every benchmark
Distribution
Where models cluster
Correlated benchmarks
Pearson r · original research
Benchmarks that track with LiveBench · Coding
Pearson correlation across models scored on both benchmarks. Closer to 1 = strongly predictive.
Full rankings
29 models tested · sorted by score
| # | Model | Score |
|---|---|---|
| 1 |  GPT-5.2-Codex | 83.6 |
| 2 | GPT-5.1-Codex-Max | 81.4 |
| 3 |  Qwen3.6 Plus | 78.2 |
| 4 | GPT-5 Mini | 76.1 |
| 5 |  DeepSeek V3.2 | 75.7 |
| 6 |  GLM 5.1 | 75.4 |
| 7 | GPT-5.4 Mini | 74.7 |
| 8 | GLM 5V Turbo | 73.9 |
| 9 | GLM 5 | 73.6 |
| 10 | DeepSeek V3.2 Exp | 73.2 |
| 11 | GLM 4.7 | 73.1 |
| 12 | GPT-5.1-Codex | 71.8 |
| 13 | GLM 4.6 | 71.0 |
| 14 |  MiniMax M2.5 | 70.7 |
| 15 | GPT-5.1-Codex-Mini | 69.9 |
| 16 | Qwen3 235B A22B Instruct 2507 | 69.6 |
| 17 | Qwen3 235B A22B Thinking 2507 | 69.0 |
| 18 |  MiMo-V2-Pro | 68.8 |
| 19 | Qwen3 Next 80B A3B Instruct | 68.2 |
| 20 |  Kimi K2 Thinking | 67.4 |
| 21 | GPT-5 Nano | 67.4 |
| 22 |  Devstral 2 2512 | 66.8 |
| 23 | GLM 4.6V | 64.2 |
| 24 | GPT-5.4 Nano | 61.9 |
| 25 | Qwen3 Next 80B A3B Thinking | 60.7 |
| 26 |  Gemma 4 31B | 60.3 |
| 27 | gpt-oss-120b | 60.2 |
| 28 | MiniMax M2.7 | 54.9 |
| 29 |  Nemotron 3 Super | 54.1 |
Frequently asked
Pulled from the LiveBench · Coding dataset · updated daily
What does LiveBench · Coding measure?
LiveBench · Coding is a knowledge benchmark in the BenchGecko catalog. 29 AI models have been tested on it. Scores range from 54.1 to 83.6 out of 100.
Which model leads on LiveBench · Coding?
GPT-5.2-Codex from OpenAI leads LiveBench · Coding with a score of 83.6. The median score across 29 tested models is 69.9.
Is LiveBench · Coding saturated?
No · the top score is 83.6 out of 100 (84%). There is still meaningful room for improvement on LiveBench · Coding.
Does LiveBench · Coding predict performance on other benchmarks?
Yes · LiveBench · Coding scores correlate 0.74 with Chatbot Arena Elo · Overall across 15 shared models. Models that do well on LiveBench · Coding tend to do well on Chatbot Arena Elo · Overall.
How often is LiveBench · Coding data refreshed?
BenchGecko pulls updates daily. New model scores on LiveBench · Coding appear as soon as they are published by Epoch AI or the model provider.
Top on LiveBench · Coding
GPT-5.2-Codex · 83.6GPT-5.1-Codex-Max · 81.4Qwen3.6 Plus · 78.2GPT-5 Mini · 76.1DeepSeek V3.2 · 75.7More knowledge benchmarks
Same category · related evaluations