The Agent Company
The Agent Company · tests AI agents on realistic corporate tasks like email management, code review, data analysis, and cross-tool workflows.
The Frontier
Best score over time · one chart, every benchmark
Full rankings
13 models tested · sorted by score
| # | Model | Score |
|---|---|---|
| 1 | 42.9 | |
| 2 | 41.1 | |
| 3 | 33.1 | |
| 4 | 30.9 | |
| 5 | 30.3 | |
| 6 | 24.0 | |
| 7 | 11.4 | |
| 8 | 8.6 | |
| 9 | 7.4 | |
| 10 | 6.9 | |
| 11 | 5.7 | |
| 12 | 3.4 | |
| 13 | 1.1 |
Score distribution
Where models cluster
Correlated benchmarks
Pearson r · original research
Benchmarks that track with The Agent Company
Pearson correlation across models scored on both benchmarks. Closer to 1 = strongly predictive.
Frequently asked
About The Agent Company
What does The Agent Company measure?
The Agent Company · tests AI agents on realistic corporate tasks like email management, code review, data analysis, and cross-tool workflows. 13 AI models have been tested on it. Scores range from 1.1 to 42.9 out of 100.
Which model leads on The Agent Company?
DeepSeek V3.2 Exp from DeepSeek leads The Agent Company with a score of 42.9. The median score across 13 tested models is 11.4.
Is The Agent Company saturated?
No · the top score is 42.9 out of 100 (43%). There is still meaningful room for improvement on The Agent Company.
Does The Agent Company predict performance on other benchmarks?
Yes · The Agent Company scores correlate 0.94 with Aider · Code Editing across 6 shared models. Models that do well on The Agent Company tend to do well on Aider · Code Editing.
How often is The Agent Company data refreshed?
BenchGecko pulls updates daily. New model scores on The Agent Company appear as soon as they are published by Epoch AI or the model provider.
- Category
- Agent
- Max score
- 100
- Models
- 13
- Updated
- 2025-09-29
Top on The Agent Company
DeepSeek V3.2 Exp · 42.9Gemini 2.5 Flash · 41.1Claude Sonnet 4 · 33.1Claude 3.7 Sonnet · 30.9Gemini 2.5 Pro · 30.3More agent benchmarks
Same category · related evaluations