GLM-5.1 is Z.ai's latest open-source model released in April 2026 under the MIT license, a post-training upgrade to GLM-5 targeting coding and agentic performance through refined reinforcement learning. Built on the same 744B MoE architecture with 40B active parameters and a 200K-token context window, it scores 58.4% on SWE-Bench Pro — surpassing Claude Opus 4.6 (57.3%) — and can autonomously manage a full plan-execute-test-fix-optimize loop for up to eight hours without human intervention. It represents one of the strongest open-weight models available for long-horizon agentic engineering tasks.
Z.aiGLM-5.1 is Z.ai's latest open-source model released in April 2026 under the MIT license, a post-training upgrade to GLM-5 targeting coding and agentic performance through refined reinforcement learning. Built on the same 744B MoE architecture with 40B active parameters and a 200K-token context window, it scores 58.4% on SWE-Bench Pro — surpassing Claude Opus 4.6 (57.3%) — and can autonomously manage a full plan-execute-test-fix-optimize loop for up to eight hours without human intervention. It represents one of the strongest open-weight models available for long-horizon agentic engineering tasks.
Z.ai