Browser Use open-sources rl-env-use: one-line RL environments for any domain
inductionheads · x · 2026-10-21
Browser Use released and open-sourced rl-env-use: describe a skill in one line and get a verified RL environment built from real web-mined cases that frontier models don't saturate, plus a hillclimbed prompt for the model you train. It builds on Anthropic's automating eval design and hillclimbing workflow, with tasks mined from the web by Browser Use agents. Benchmarks: on 8 clinical-trial-result missions Claude Code verified 0 of 86 facts while Browser Use verified 17 with 4 complete cases; on human genetic variants (UniProt, gnomAD, GTEx) it's 6 of 82 vs 40 verified.
More from coding & agent
- sentence-transformers v6.2.0 ships model merging, faster loading, distributed eval across 176 PRs — tomaarsen · 2026-10-21
- Sentence Transformers 6.2.0 adds Python 3.14 support with two upgrade caveats — tomaarsen · 2026-10-21
- Sentence Transformers 6.2.0 fixes FP16 pooling arithmetic, speeds up CPU semantic search — tomaarsen · 2026-10-21
- Sentence Transformers 6.2.0 improves multi-vector models with finer compression control — tomaarsen · 2026-10-21
- Sentence Transformers 6.2.0: DDP eval sharing and 43-61% faster cached model loading — tomaarsen · 2026-10-21
- Sentence Transformers 6.2.0 cuts training memory with quadratic/linear loss storage — tomaarsen · 2026-10-21