Model releases  ·  11 August 2026

GLM-5.2: MIT-licensed flagship built for long-horizon work

Z.ai's current flagship pairs a million-token context with a sparse attention scheme designed specifically to make that length affordable — and ships it with no regional restrictions attached.

At a glance

Released byZ.ai
PublishedJune 2026, last updated July 2026
ArchitectureMixture-of-Experts with sparse attention (GlmMoeDsaForCausalLM)
Layers / hidden size78 / 6144
Experts256 routed, top-8 per token, plus 1 shared
Context length1,048,576 tokens (1M)
Vocabulary154,880 tokens
LicenceMIT
Weight formatsafetensors — 1,506.7 GB across 282 shards
Also publishedGLM-5.2-FP8 quantised build
Hosted atzai-org/GLM-5.2 on Hugging Face

What it is

GLM-5.2 is Z.ai's flagship model for long-horizon tasks — work that runs across many steps rather than single exchanges. Z.ai frames it as a substantial step over GLM-5.1 and, for the first time in the family, one that delivers on what they call a "solid" 1M-token context: not merely accepting a million tokens, but sustaining useful work across them.

IndexShare

The architectural claim is specific and measurable. IndexShare reuses the same indexer across every four sparse attention layers, which Z.ai reports as a 2.9× reduction in per-token FLOPs at a 1M-token context length. They also report improving the multi-token-prediction layer used for speculative decoding, raising acceptance length by up to 20%.

The model type string, glm_moe_dsa, reflects the sparse attention design rather than a conventional dense-attention MoE.

The licence is deliberate positioning

Z.ai's own release note calls it "Pure Open: an MIT open-source license — no regional limits, technical access without borders." That is unusually direct language, and it puts GLM-5.2 alongside Motif 3 and DeepSeek V4 Flash in the group of very large models with no acceptable-use appendix attached.

What's in the release

The main repository is 1,506.7 GB across 282 safetensors shards. Z.ai also publishes GLM-5.2-FP8, a quantised build that has drawn more than two million downloads on its own — a reasonable indication that most real deployments start there rather than with the full-precision weights.

The wider family on the same account includes GLM-5, GLM-5.1, GLM-4.7 and GLM-4.7-Flash, plus specialised GLM-OCR and GLM-ASR models, all MIT.

Runtime support

RuntimeSupportedNotes
transformersYesStandard configuration published by the lab.
Official quantised buildYesGLM-5.2-FP8 is published by Z.ai alongside the full-precision weights.
llama.cppNoglm_moe_dsa sparse attention has no announced upstream support.
MLXNoNo conversion published by the lab.

What we don't know yet

Sources

Frequently asked questions

What licence is GLM-5.2 released under?

MIT, which Z.ai describes as carrying no regional limits and no additional access restrictions.

What is IndexShare?

A sparse attention design that reuses the same indexer across every four attention layers. Z.ai reports it reduces per-token FLOPs by 2.9 times at a 1M-token context length.

Is there an official quantised GLM-5.2?

Yes. GLM-5.2-FP8 is published by Z.ai and has more downloads than the full-precision repository.

How large is GLM-5.2?

1,506.7 GB across 282 safetensors shards for the full-precision weights. The FP8 build is substantially smaller.

Can llama.cpp run GLM-5.2?

Not as of publication. The glm_moe_dsa sparse attention architecture has no announced upstream support.

OnDevice LLM is a private AI assistant that runs entirely on your iPhone — no account, no cloud, and nothing you type leaves the device.

Published 11 August 2026