ProgramDistill: From Interactive Web Apps to Verifiable Reference-Guided SWE Tasks
Abstract
Coding agents are typically evaluated with desired behavior specified through issues or instructions. In practical web development, however, agents may need to infer behavior from working software and implement it in an incomplete application. We introduce ProgramDistill, a benchmark evaluating coding agents on features discovered through interaction with fully functional reference applications. We build ProgramDistill by factorizing applications into features of different granularities, each associated with replayable behaviors executable via its gold patch. Our pipeline, mine-craft-patch, discovers 1,975 replay-verified behaviors across 26 applications and constructs 4,063 tasks without human intervention. Across nine frontier coding agents, GPT-6 Astra and Claude Opus 5 achieve 49.2% and 28.8% success on cumulative workflows in full-application reconstruction. In partial-application reconstruction, success falls from 100% to 64.0% and from 96% to 32% as restoration depth increases from 1 to 8. ProgramDistill thus provides a scalable benchmark with controlled difficulty for evaluating and diagnosing coding agents, and a natural basis for future curriculum-based training.
Community
For more information, please check out our:
📝 Blog: https://microsoft.github.io/debug-gym/blog/2026/09/programdistill/
🔗 Paper: https://microsoft.github.io/debug-gym/static/papers/ProgramDistill_arxiv.pdf
We’re working on a public release of the ProgramDistill benchmark, so you'll be able to play with it yourself soon.
Get this paper in your agent:
hf papers read 2609.18805 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper