MyPCBench: leaderboard
Personal-computer task benchmark evaluating computer-use agents on realistic desktop workflows with screenshots, browser/filesystem actions, shell access, and rubric-based grading.
Metric: Perfect Rate (%). Source: mypcbench.com. Saturation forecast: Around January 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 58.2 | #63 |
| 2 | GPT-5.6 Sol | 55.4 | #19 |
| 3 | GPT-5.6 Luna | 55.4 | #68 |
| 4 | Claude Sonnet 4.6 | 50.5 | #90 |
| 5 | GPT-5.5 | 45.1 | #31 |
| 6 | GPT-5.4 Mini | 23.9 | #212 |
| 7 | Qwen 3.5 35B A3B | 7.6 | #256 |
| 8 | Qwen 3.5 9B | 6.5 | #379 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mypcbench · How It Works · Data refreshed daily, snapshot 2026-10-20.