GLM-5.3

Specs

Input price

$1.4/M

Output price

$4.4/M

Context window

1M tokens

Capabilities

• 753B parameter MoE with every gain from a month of post-training
• Terminal-Bench 2.1 at 88.2 - ahead of DS-V4-Pro-0813 and Opus-4.8
• First open model over the frontier tier on Terminal-Bench 3.0 (28.3)
• Frontier-level defensive security: CyberGym 84.5, ahead of Opus-4.8

Best for

• Long-horizon agentic coding and automation (AutomationBench 26.2 to 48.2)
• Repository-scale engineering over a 1M-token context window
• Defensive security workflows: vulnerability discovery and review automation

Limitations to keep in mind

• Text-only - vision work rides on GLM-5.3-Flash or another vision model
• Custom glm-5.3 license: MaaS operators over $10B revenue need a Z.ai security review
• No new pre-training run - conservative architecture, post-training gains only

HIPAA-compliant hosting

GLM-5.3 is available under our HIPAA-compliant design-partner program, with a signed Business Associate Agreement (BAA), encryption in transit and at rest, and access controls. We are currently onboarding design partners, with general availability coming soon. Run the model behind a unified OpenRouter-style API and swap to another model with a single parameter.

Pricing in context

Open source models like GLM-5.3 can be served efficiently on optimized inference infrastructure, with savings passed through to you. Exact savings depend on the model and your volume, but open source inference is typically a fraction of the per-token cost of closed-source frontier models on Azure OpenAI or AWS Bedrock - without cloud egress lock-in or minimum commitments.