Doc2Harness: Benchmarking LLMs on Long-Horizon Harness Engineering of Coding Agents Paper • 2610.20344 • Published 9 days ago • 63
AgentSWE: Can Coding Agents Build the Agent You Actually Want? Paper • 2610.13361 • Published 14 days ago • 93
LLaDA-Image: Building Strong Image Generators with Fully Open Training Recipes Paper • 2609.03796 • Published Sep 3 • 113