Environment synthesis

Technique · Agents · 3 items since Jun 2026

Automatically building executable environments for training agents with reinforcement learning.

Trends

Sep 4
Qwen's Terminal-Universe rebuilds agent workspaces from trajectories
Instead of building agent-training environments from scratch, the framework replays tool-execution histories to reconstruct 37.3k executable workspaces, lifting Qwen3.5-27B by 11.9 points on Terminal-Bench 2.1 and 13.8 on multi-round EvoCode-Bench v2 after fine-tuning.
Aug 15
Google's EnvHarness reshapes agent training environments
A programmable plug-in layer reshapes static agent environments without changing their logic or verifier, while an automated rigger synthesizes components that target a policy's diagnosed flaws, yielding up to 9.0-point gains with 9.8% fewer execution steps.
Jun 2
GAIS turns MCP servers into training environments
Grounded Interaction Synthesis converts real Model Context Protocol servers into executable Python environments and plans tasks around deep inter-tool dependencies, letting base Qwen3 models trained on the data match or beat their official instruction-tuned versions on BFCL, τ²-Bench, and ACEBench.