GLM-5.1 is a new open-source flagship AI model from ZhipuAI focused on long-horizon agentic engineering tasks. Unlike previous models that plateau quickly, GLM-5.1 sustains productive optimization over hundreds of iterations and thousands of tool calls. Key demonstrations include: a vector search optimization task (VectorDBBench) where it reached 21.5k QPS over 600+ iterations vs. 3,547 QPS in a single 50-turn session; GPU kernel optimization (KernelBench Level 3) achieving 3.6× speedup; and an open-ended web app build producing a full Linux-style desktop environment in the browser after 8 hours of self-directed iteration. The model achieves state-of-the-art on SWE-Bench Pro (58.4) and leads competitors on NL2Repo and Terminal-Bench 2.0. It is released under the MIT License with weights on HuggingFace and ModelScope, and is compatible with Claude Code, Cline, and other coding agents.
Sort: