Google open-sources Harness², allowing AI agents to self-evolve and improve without changing model weights
Google開源Harness 2，不改模型權重AI代理人可自我演進提升
A Google research team open-sourced the Harness² framework under the Apache 2.0 license, letting AI agents dynamically adjust and self-evolve based on experience without modifying the model's underlying weights. The framework is implemented through two levels of recursion: the task-level recursion targets a single task, and the editor proposes candidate solutions in three steps—Propose, Probe, Compose—and retains improvements verified by behavior, without needing ground-truth feedback; the Harnessing-level recursion has a Reflector record adopted modifications and lessons learned, forming an experience library to guide subsequent tasks. Across three professional benchmarks and 4 model harness configurations, a single Harness² step improved the task performance of GLM-5 and Gemini 3.5 Flash, with WorkBuddy-Bench improving by up to 11.5 points, JobBench by up to 4.5 points, and LAB partial pass rate increasing by up to 11.8 percentage points.
Google's open-source Agent framework demonstrates a path to improving task success rates without changing model weights, relying only on an external framework, and includes measured improvement data on three benchmarks.