
GLM-5.3
GLM-5.3 is Z.ai's open-weight model for coding and long-horizon agentic tasks. It uses the same base model as GLM-5.2, and its gains come from post-training.
Ways to run GLM-5.3
Dedicated endpoint
Get a quote →Your own GPUs, tuned for your traffic. We profile your workload, tune the deployment and put throughput and latency in your contract.
Serverless
API docs →Pay per token on shared GPUs, with no capacity to plan or reserve.
Inside your network
Security →Run a dedicated endpoint inside your own network, including air-gapped environments.
Pricing
2.4x the throughput after tuning
Total tokens per second per GPU, 16 live agent sessions, same four GB300 GPUs
Tuning added decode context parallel 4, a 256 GB CPU KV cache tier and multi-token prediction with 3 draft tokens. Your numbers will differ with your prompts and traffic.
How we measured thisUse it with the agents we set up for you
Point any open-source agent at your endpoint, or pick from our catalog and we set it up as part of your contract.
GLM-5.3 FAQ
How much does GLM-5.3 cost on Saturn Cloud?
Serverless costs $0.35 per 1M input tokens and $1.10 per 1M output tokens. A dedicated endpoint is priced per GPU-hour for the GPUs it runs on, and we size the GPUs to your traffic.
What did tuning change for GLM-5.3?
We added decode context parallel 4, a 256 GB CPU KV cache tier and multi-token prediction with 3 draft tokens. The starting setup was our catalog defaults on one endpoint.
What is GLM-5.3's context length?
The Hugging Face config sets 1M tokens, and Z.ai evaluates agents at up to 1M. In our four-GPU benchmark setup, the context limit was 524,288 tokens.
How do I control reasoning in GLM-5.3?
Set reasoning_effort to low, high or max. The default is max. For chat, pass clear_thinking=true, since it defaults to false.
How does GLM-5.3 compare with GLM-4.6?
GLM-5.3 is Z.ai’s newer model, with a 1M-token context window against GLM-4.6’s 200K. On serverless, GLM-5.3 costs $0.35 input and $1.10 output per 1M tokens, and GLM-4.6 costs $0.075 and $0.30.