Meta新出的Harness-of-Harness能让编程代理连续工作数天，比独立系统平均提升52%，最高82%
Meta推出的Harness-of-Harness系统在GameCraft-Bench、FrontierSWE和ProgramBench三个基准测试中，与三种不同的harness和模型配对相比，经过三次迭代后平均提升52.25%，最高达82.86%。该系统通过将执行组织为规划、编码和测试的重复增量，平衡修复与能力增长，将工作范围限定为可验证的小步骤。系统保持实现时测试与独立评估分离，约束输出而非工作流程。
Meta harnesses are fascinating. Pay attention if you are an AI engineer. A different problem set, bu...
Meta harnesses are fascinating. Pay attention if you are an AI engineer. A different problem set, but ridiculously powerful if done right. I'm interested in emerging properties like better token efficiency, orchestration, routing, new test-time compute scaling... DAIR.AI @dair_ai // Harness-of-Harness // Exciting new research on coding agents that keep building for days without a human. Here is how it works: Harness-of-Harness wraps whatever coding harness you already run and organizes its executions into repeated planning, coding and testing increments. The loop balances repair against capability growth, scopes work into small verifiable steps, keeps implementation-time testing separate from independent evaluation, and constrains the outputs rather than the workflow. Across GameCraft-Bench, FrontierSWE and ProgramBench with three different harness and model pairs, it averages a 52.25 percent relative gain over the standalone harnesses after three iterations, peaking at 82.86 percent. Paper: arxiv.org/abs/2609.01481 Chat with Paper: academy.dair.ai/papers/harness… 🔗 View Quoted Tweet 💬 6 🔄 2 ❤️ 27 👀 3438 📊 11 ⚡