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

arXiv: https://arxiv.org/abs/2607.28074 · HTML: https://mlsanity.com/paper/2607.28074/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use.
- Builds on 25 of its 35 references that are in the map.
- First posted on arXiv 30 Jul 2026.

## Placed in

- LLM agents and tool use / [General LLM agent frameworks and benchmarks](https://mlsanity.com/agents/general-llm-agent-frameworks-and-benchmarks/)

## Builds on (most cited first, 25 of 25)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains](https://mlsanity.com/paper/2406.12045/) (2024)
- [Understanding R1-Zero-Like Training: A Critical Perspective](https://mlsanity.com/paper/2503.20783/) (2025)
- [Mind2Web: Towards a Generalist Agent for the Web](https://mlsanity.com/paper/2306.06070/) (2023)
- [OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments](https://mlsanity.com/paper/2404.07972/) (2024)
- [STaR: Bootstrapping Reasoning With Reasoning](https://mlsanity.com/paper/2203.14465/) (2022)
- [AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents](https://mlsanity.com/paper/2405.14573/) (2024)
- [WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models](https://mlsanity.com/paper/2401.13919/) (2024)
- [Scaling Relationship on Learning Mathematical Reasoning with Large Language Models](https://mlsanity.com/paper/2308.01825/) (2023)
- [WorkArena: How Capable Are Web Agents at Solving Common Knowledge Work Tasks?](https://mlsanity.com/paper/2403.07718/) (2024)
- [Emergent Complexity and Zero-shot Transfer via Unsupervised Environment Design](https://mlsanity.com/paper/2012.02096/) (2020)
- [Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions](https://mlsanity.com/paper/1901.01753/) (2019)
- [An Illusion of Progress? Assessing the Current State of Web Agents](https://mlsanity.com/paper/2504.01382/) (2025)
- [Evolving Curricula with Regret-Based Environment Design](https://mlsanity.com/paper/2203.01302/) (2022)
- [The BrowserGym Ecosystem for Web Agent Research](https://mlsanity.com/paper/2412.05467/) (2024)
- [Automatic Curriculum Learning For Deep RL: A Short Survey](https://mlsanity.com/paper/2003.04664/) (2020)
- [Fara-7B: An Efficient Agentic Model for Computer Use](https://mlsanity.com/paper/2511.19663/) (2025)
- [UltraCUA: A Foundation Model for Computer Use Agents with Hybrid Action](https://mlsanity.com/paper/2510.17790/) (2025)
- [InfiniteWeb: Scalable Web Environment Synthesis for GUI Agent Training](https://mlsanity.com/paper/2601.04126/) (2026)
- [Gym-Anything: Turn any Software into an Agent Environment](https://mlsanity.com/paper/2604.06126/) (2026)
- [CUA-Gym: Scaling Verifiable Training Environments and Tasks for Computer-Use Agents](https://mlsanity.com/paper/2605.25624/) (2026)
- [The Art of Building Verifiers for Computer Use Agents](https://mlsanity.com/paper/2604.06240/) (2026)
- [Fara-1.5: Scalable Learning Environments for Computer Use Agents](https://mlsanity.com/paper/2606.20785/) (2026)

## Built on by (most cited first, 2 of 2)

- [WorkGenesis: Building the Worlds That Teach Agents to Work](https://mlsanity.com/paper/2609.39325/) (2026)
- [WorkRover: Scaling Long-Horizon Cowork Agent Training Via Recursive Environment Evolution](https://mlsanity.com/paper/2610.18067/) (2026)

## Abstract (arXiv)

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 → 75.0) while deep ones raise it (80.0 → 85.0 and 48.0 → 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