GLM 5.3
Chat model · Z.ai ·
GLM 5.3 is a chat model by Z.ai. In Endless it costs 113.4 credits per 1M input tokens and 356.4 per 1M output tokens, on every plan.
GLM 5.3 is Z.ai's flagship model for complex software engineering and multi-step agent work, reasoning through every request over a 1M-token context.
- Price
- 113.4 credits per 1M input tokens
- 356.4 credits per 1M output tokens
- Availability
- Available on every plan
- Context window
- 1M tokens
- Max output
- 128K tokens
- Inputs
- Text
- Outputs
- Text
About GLM 5.3
GLM 5.3 is Z.ai's flagship model. It keeps the GLM-5.2 base and gets all of its gains from more post-training, aimed at software engineering and agent work. Reasoning is always on: you choose low, high or max effort, and max is the right setting for coding and other complex tasks. It reads text only and holds up to 1M tokens of context.
Use it when a task has many dependent steps: tracing a bug across services, planning and executing a migration, or working through a terminal session where each command depends on the last. Its post-training also targeted defensive security work, such as reviewing code for vulnerabilities.
GLM 5.2 runs on the same base model and lets you turn reasoning off for quick replies. GLM 5.3 is the choice when accuracy on multi-step work matters more than speed. Both are available in Endless chat, and you can switch between them within one workspace.
Best for
- Multi-step debugging across a large codebase
- Planning and carrying out code migrations
- Terminal and agent workflows with many tool calls
- Reviewing code for security issues
Questions about GLM 5.3
- What changed between GLM 5.2 and GLM 5.3?
- GLM 5.3 keeps the same base model and improves on GLM 5.2 entirely through post-training, focusing on software engineering, agent tasks and defensive security. Reasoning can no longer be switched off.
- Can I turn off thinking in GLM 5.3?
- No. GLM 5.3 always reasons. You can lower the effort to low for lighter tasks, or keep it at max, the better setting for coding.
- Does GLM 5.3 accept images?
- No. GLM 5.3 currently takes text only, so describe visuals in words or pick a model with image input for screenshots.
Related models
GLM 5.2
by Z.ai
Project-scale coding and tool use over a 1M-token context
64.8 / 206.6 credits per 1M tokens
GPT-6.1 Sol
by OpenAI
Near-flagship agentic coding and professional work at lower cost
162 / 810 credits per 1M tokens
Claude Opus 5.5
by Anthropic
Long agentic coding and research work over a 1M-token context
324 / 1620 credits per 1M tokens
Give your next idea a place to grow.
Start free, create a project, and bring your brand’s references together. Or book a demo to explore Endless with your team.