~753.9B total (MoE-DSA family; FP8 variant also released)
202,752 tokens
MIT
Open-weight GLM-5.1 release positioned as Z.AI's next-generation flagship for agentic engineering, with official materials claiming significantly stronger coding capability than GLM-5. The model card reports state-of-the-art performance on SWE-Bench Pro and sizable gains over GLM-5 on NL2Repo and Terminal-Bench 2.0, while also improving across broader tool-use and web-browsing evaluations in the published benchmark table. Release messaging emphasizes long-horizon agent reliability rather than only first-pass quality, claiming GLM-5.1 can sustain iterative optimization over hundreds of rounds and thousands of tool calls by revising strategy based on intermediate results. The release includes both a standard BF16 checkpoint and an FP8 variant, with local serving guidance across SGLang, vLLM, xLLM, Transformers, and KTransformers.