LiveBench · Agentic Coding
The Frontier
Best score over time · one chart, every benchmark
Full rankings
40 models tested · sorted by score
| # | Model | Score |
|---|---|---|
| 1 | 73.3 | |
| 2 | 70.0 | |
| 3 | 60.0 | |
| 4 | 56.7 | |
| 5 | 56.7 | |
| 6 | 55.0 | |
| 7 | 55.0 | |
| 8 | 55.0 | |
| 9 | 53.3 | |
| 10 | 53.3 | |
| 11 | 51.7 | |
| 12 | 51.7 | |
| 13 | 51.7 | |
| 14 | 50.0 | |
| 15 | 50.0 | |
| 16 | 50.0 | |
| 17 | 50.0 | |
| 18 | 46.7 | |
| 19 | 46.7 | |
| 20 | 46.7 | |
| 21 | 43.3 | |
| 22 | 41.7 | |
| 23 | 40.0 | |
| 24 | 40.0 | |
| 25 | 38.3 | |
| 26 | 36.7 | |
| 27 | 35.0 | |
| 28 | 35.0 | |
| 29 | 30.0 | |
| 30 | 28.3 | |
| 31 | 27.1 | |
| 32 | 23.0 | |
| 33 | 17.0 | |
| 34 | 16.7 | |
| 35 | 13.3 | |
| 36 | 10.0 | |
| 37 | 8.3 | |
| 38 | 6.7 | |
| 39 | 3.3 | |
| 40 | 3.3 |
Score distribution
Where models cluster
Correlated benchmarks
Pearson r · original research
Benchmarks that track with LiveBench · Agentic Coding
Pearson correlation across models scored on both benchmarks. Closer to 1 = strongly predictive.
Frequently asked
About LiveBench · Agentic Coding
What does LiveBench · Agentic Coding measure?
LiveBench · Agentic Coding is a knowledge benchmark in the BenchGecko catalog. 40 AI models have been tested on it. Scores range from 3.3 to 73.3 out of 100.
Which model leads on LiveBench · Agentic Coding?
GLM 5.2 from z-ai leads LiveBench · Agentic Coding with a score of 73.3. The median score across 40 tested models is 45.0.
Is LiveBench · Agentic Coding saturated?
No · the top score is 73.3 out of 100 (73%). There is still meaningful room for improvement on LiveBench · Agentic Coding.
Does LiveBench · Agentic Coding predict performance on other benchmarks?
Yes · LiveBench · Agentic Coding scores correlate 0.93 with Terminal Bench across 9 shared models. Models that do well on LiveBench · Agentic Coding tend to do well on Terminal Bench.
How often is LiveBench · Agentic Coding data refreshed?
BenchGecko pulls updates daily. New model scores on LiveBench · Agentic Coding appear as soon as they are published by Epoch AI or the model provider.
- Category
- Knowledge
- Max score
- 100
- Models
- 40
- Updated
- 2026-06-16
Top on LiveBench · Agentic Coding
GLM 5.2 · 73.3Kimi K2.7 Code · 70.0MiniMax M3 · 60.0DeepSeek V4 Pro · 56.7GPT-5.1-Codex-Max · 56.7More knowledge benchmarks
Same category · related evaluations