Microsoft's Echoverse: Deep Evolving Environments Boost Computer-Use Agent from 36.5% to 67.1%
MicrosoftResearch · hf · 2026-07-31
Microsoft Research presents Echoverse, a deep evolving environment for training computer-use agents. It compiles specs into stateful apps with database-graded tasks and a co-evolution loop that uses graded rollouts for repairs and training. A 9B model improves from 36.5% to 67.1% across 14 splits, within 14 points of its frontier teacher. Shallow environments hurt live-site accuracy (80.0 to 75.0) while deep ones help (80.0 to 85.0); repairing one environment lifts score from 16.2% to 38.5%.
More from coding & agent
- 'LLM code is sloppy' rebutted with real ffmpeg code: identical if/else branches in a top library — kingslyj · 2026-10-16
- Agent actions need checks outside the model: NVIDIA demos lean on guardrails amid wave of real agent incidents — Creamy-And-Crowded · 2026-10-16
- Vitalik builds "paranoid mode" agent wrapper with Tor, zkAPI and local prompt rewriting — DavideCrapis · 2026-10-16
- Agent fine-tunes Qwen3.5-0.8B with synthetic data: 4.7%→76.3% accuracy, 11.9× faster — ben_burtenshaw · 2026-10-16
- Beam-501B agent demo: solo beats 5-agent team on ProgramBench rebuild task — ben_burtenshaw · 2026-10-16
- Beam agent builds BeamSpeak, a local Parakeet dictation app for macOS — ben_burtenshaw · 2026-10-16