Echoverse: Deep, Evolving Environments for Training Computer-Use Agents at Scale

2026-07-30

AI Summary

Current agent training lacks realistic synthetic environments that are executable, resettable, and exhibit behavioral depth—particularly in login-restricted and state-dependent scenarios. This work proposes a framework for specifying environments that compile into stateful applications, enabling co-evolutionary training wherein evaluation trajectories simultaneously inform environment refinement and agent optimization. The framework integrates a database-driven task scorer, dense step-level rewards, and a grounding validation mechanism. Experiments demonstrate that a 9B-parameter model improves from 36.5% to 67.1% average performance across 14 benchmarks, approaching the capabilities of much larger state-of-the-art models. A single round of environment repair doubles performance, and under reinforcement learning settings, the agent achieves 68.0% success, substantially enhancing online accuracy.

Abstract

Computer-use agents learn from what their actions change, so training one needs applications it can act on, break and reset. The applications that matter most are login-gated and stateful, so synthetic environments stand in for them. Recent pipelines generate such environments in bulk, which moves the bottleneck from how many exist to what is inside each one. The returns, we find, come from three properties: how much behavioural depth an environment carries, whether it targets the interaction an agent actually fails, and whether it improves alongside the model. We present Echoverse, which compiles specifications into stateful applications whose tasks are graded against the application's own database, and a co-evolution loop that reads every graded rollout twice: as repairs to the environment, its tasks and its verifier, and as training signal for the model. Trained on twelve such environments, a 9B model improves from $36.5\%$ to $67.1\%$ across fourteen evaluation splits, within fourteen points of the much larger frontier model that taught it. We examine each property in turn. On the same domains, shallow environments push live-site accuracy below the base model ($80.0 \to 75.0$) while deep ones raise it ($80.0 \to 85.0$ and $48.0 \to 65.0$); drilling one interface control across many renderings transfers to held-out widget families and to the open web; and repairing a single environment lifts the model trained on it from $16.2\%$ to $38.5\%$. The same worlds serve as reinforcement-learning environments, where a reward combining the grounded verifier with a dense per-step judge raises held-out score from $58.8\%$ to $68.0\%$. We release four environments as a benchmark, with their applications, seed data and grounded graders. Code: https://aka.ms/echoverse

Problem

• computer-use agents
• synthetic environments
• behavioral depth
• co-evolution
• stateful applications

Innovation

• co-evolution
• behavioural depth
• stateful environments
• grounded verification
• computer-use agents