calcom/
Scheduling infrastructure for absolutely everyone.
Run by mupt-ai · Released 2026-10-26 · 39 tasks · 7 settings
Which coding agent works best on calcom/cal.diy? Most accurate: Claude Opus 5.5, 74.4% at $6.39 per task. Best for less: Dari Router, 69.2% at $0.22. 7 model settings scored on 39 tasks from its merged pull requests.
All Settings
| Model | Harness | Reasoning | Access | Accuracy | Cost / Task | Frontier |
|---|---|---|---|---|---|---|
| Claude Opus 5.5 | Claude Code | Max | API Key | 74.4% | $6.39 | Yes |
| Dari Router | Pi | Default | API Key | 69.2% | $0.22 | Yes |
| GPT-6.1 Sol | Codex | Medium | API Key | 69.2% | $0.33 | No |
| GPT-6.1 Sol | Codex | Max | API Key | 69.2% | $0.56 | No |
| DeepSeek V4.1 Flash 1017 | Pi | High | Vercel AI Gateway | 61.5% | $0.072 | Yes |
| GPT-6 Luna | Codex | Max | API Key | 53.8% | $0.047 | Yes |
| GPT-6 Luna | Codex | Medium | API Key | 48.7% | $0.015 | Yes |
Results by Task
Whether each setting passed each task, with how its attempt was graded and its transcript.
Tasks
All 39 tasks, each with its instruction, tests, and solution to browse and download.