Model library / zai-org
GLM-5.3 is a 753.3 billion-parameter open-weight mixture-of-experts language model for coding and security tasks.
8 of 256 experts active per token
Context length in tokens, as published
Weights at the released precision
Licence declared on the repository
Overview
GLM-5.3 is a 753.3 billion-parameter mixture-of-experts language model built on the GLM-5.2 base. The publisher, zai-org, describes it as an open-weight model aimed at complex coding and long-horizon tasks, with post-training improvements for coding, cyber security and exploitation benchmarks. It targets developers and security researchers who need high-capacity reasoning over extended contexts.
Deploying GLM-5.3 requires handling a 753.3 billion-parameter model with 256 experts, eight active per token, and fp8 weights. It supports a context window of up to 1,048,576 tokens and is released under a non-standard licence, so users must check compatibility before integration. The model can be served via frameworks such as SGLang, vLLM, TokenSpeed, Transformers, KTransformers and Unsloth, and runs on Ascend NPU platforms with appropriate back-ends.
Specification
| Repository | zai-org/GLM-5.3 |
| Publisher | zai-org |
| Published | 25 August 2026 |
| Architecture | GlmMoeDsaForCausalLM |
| Model type | glm_moe_dsa |
| Parameters | 753,329,940,480 |
| Experts | 256 total, 8 active per token |
| Layers | 78 |
| Hidden size | 6,144 |
| Attention heads | 64, 64 key/value heads |
| Context length | 1,048,576 tokens |
| Vocabulary | 154,880 tokens |
| Quantisation | fp8 |
| Licence | other |
Memory
Parameter count multiplied by the bytes each format uses. The key-value cache comes on top.
| Precision | Weights | Notes |
|---|---|---|
| bf16 | 1403.2 GiB | Full precision as released |
| fp8 | 701.6 GiB | 8-bit, near-lossless on most models |
| int4 | 350.8 GiB | 4-bit, smallest footprint |
The cache costs 3.7 MiB per token at 16-bit, so the full 1,048,576-token context of one request needs about 3744.0 GiB. Concurrency multiplies that number, not the weights.
That figure is an upper bound: this model uses compressed key-value attention, so the real cache is a fraction of the figure above.
Hardware
At the released precision, with 90% of the card given to the server and 2.5 GiB kept for the runtime. The context column is what is left for the cache on that setup.
| GPU | Cards needed | Free for cache | Context it holds |
|---|---|---|---|
| RTX 4090 24GB | more than 8 | — | — |
| RTX 5090 32GB | more than 8 | — | — |
| L40S 48GB | more than 8 | — | — |
| A100 80GB | more than 8 | — | — |
| H100 80GB | more than 8 | — | — |
| RTX PRO 6000 96GB | more than 8 | — | — |
| H200 141GB | 8 | 291.6 GiB | about 81,681 tokens |
Serving
A vLLM launch line for the shape above. Check the model card for a runtime the publisher recommends.
vllm serve zai-org/GLM-5.3 \ --tensor-parallel-size 8 \ --max-model-len 77824 \ --gpu-memory-utilization 0.90
Questions
The weights take about 704 GiB at the released precision. No single card in the table above holds that, so it needs several GPUs or a lower precision.
The repository declares other. Read the licence text before commercial use: the name alone does not tell you what is allowed.
The configuration allows 1,048,576 tokens. Whether the whole window is usable depends on the memory left for the cache, which the hardware table works out per card.
This page is a reference, and Aiqre does not serve this model today. We deploy open-weight models on dedicated EU hardware on request, so if you want an OpenAI-compatible endpoint for it, email [email protected] with the model name and roughly what volume you expect.
Also in the library
Nex-N2.5-Pro is a 396.8 billion-parameter MoE language model designed for long-horizon, agentic tasks with multimodal and compute…
GigaChat 3.5 Reasoning is a 438 billion-parameter mixture-of-experts language model trained with online RL for improved reasoning…
K2-Horizon-375B-A23B is an open-weight sparse MiE language model from IFM with 379 B parameters, 23 B active per token and a 512 …
Hy4 preview is a 780 billion-parameter mixture-of-experts language model released by Tencent for productivity-focused tasks.
Hosting
Tell us the model and roughly what volume you expect. We reply with a price and an OpenAI-compatible endpoint, with your prompts never stored.