GLM-5

GLM-5.1 - Overview - Z.AI DEVELOPER DOCUMENT

Navigation Language Models GLM-5.1 Guides API Reference Scenario Example Coding Plan Released Notes Terms and Policy Help Center Get Started Quick Start Overview Pricing Core Parameters SDKs Guide Migrate to GLM-5.1 Language Models GLM-5.1 GLM-5 GLM-5-Turbo GLM-4.7 GLM-4.6 GLM-4.5 GLM-4-32B-0414-128K Vision Language Models GLM-5V-Turbo GLM-4.6V GLM-OCR AutoGLM-Phone-Multilingual GLM-4.5V Image Generation Models GLM-Image CogView-4 Video Generation Models CogVideoX-3 Vidu Q1 V

Scaling laws push model capability forward. But whether that capability becomes reliable in production depends on how we handle Scaling Pain. https://t.co/o0k0E0hOAp In our latest blog, we share how w

Scaling laws push model capability forward. But whether that capability becomes reliable in production depends on how we handle Scaling Pain. https://t.co/o0k0E0hOAp In our latest blog, we share how we debugged GLM-5 serving at scale: reproducing rare garbled outputs,

MISA: Mixture of Indexer Sparse Attention for Long-Context LLM Inference

DeepSeek Sparse Attention (DSA) sets the state of the art for fine-grained inference-time sparse attention by introducing a learned token-wise indexer that scores every prefix token and selects the most relevant ones for the main attention. To remain expressive, the indexer uses many query heads (for example, 64 on DeepSeek-V3.2) that share the same selected token set; this multi-head design is precisely what makes the indexer the dominant cost on long contexts. We propose MISA (Mixture of Indexer Sparse Attention), a drop-in replacement for the DSA indexer that treats its indexer heads as a pool of mixture-of-experts. A lightweight router uses cheap block-level statistics to pick a query-dependent subset of only a few active heads, and only those heads run the heavy token-level scoring. This preserves the diversity of the original indexer pool while reducing the per-query cost from scoring every prefix token with every head to scoring it with only a handful of routed heads, plus a negligible router term computed on a small set of pooled keys. We further introduce a hierarchical variant of MISA that uses the routed pass to keep an enlarged candidate set and then re-ranks it with the original DSA indexer to recover the final selected tokens almost exactly. With only eight active heads and no additional training, MISA matches the dense DSA indexer on LongBench across DeepSeek-V3.2 and GLM-5 while running with eight and four times fewer indexer heads respectively, and outperforms HISA on average. It also preserves fully green Needle-in-a-Haystack heatmaps up to a 128K-token context and recovers more than 92% of the tokens selected by the DSA indexer per layer. Our TileLang kernel delivers roughly a 3.82 times speedup over DSA's original indexer kernel on a single NVIDIA H200 GPU.

GLM-5 is now available through Ollama Cloud. 198K context window listed. A strong reasoning and agentic model from Z.ai with 744B total parameters (40B active), built for complex systems engineering and long-horizon tasks.