How do I use GLM 5.3 in Codex CLI?
Run standardcompute launch codex -- --model z-ai/glm-5.3, or add a Responses provider at https://api.stdcmpt.com/v1 with model = "z-ai/glm-5.3": GLM 5.3 then runs Codex on a plan from $19 a month that stops new requests with HTTP 402 once the compute budget is used. Z.ai's own docs list no Responses API, which current Codex releases require.
# Install Codex and the Standard Compute launcher once
npm install -g @openai/codex
curl -fsSL https://standardcompute.com/cli/install.sh | sh
# Start Codex pinned to z-ai/glm-5.3
standardcompute launch codex -- --model z-ai/glm-5.3
Swap in z-ai/glm-5.3-flash for the lower-priced sibling that also reads images.
GLM 5.3 or GLM 5.3 Flash in Codex
| z-ai/glm-5.3 | z-ai/glm-5.3-flash | |
|---|---|---|
| Images | Text-only: an image fails with HTTP 400 | Reads images |
| Typical day at Z.ai list price | $3.65 | $0.41 |
| Busy day | $37 | $4.14 |
| Typical days in the $20 Starter budget | 5 | 49 |
Z.ai list prices per million tokens: GLM 5.3 $1.40 input, $0.26 cached, $4.40 output; Flash $0.15, $0.03, $0.50. Source. A typical day is 8.5M tokens and a busy one 86.7M, 87.2% of input as cache reads: the median and 90th-percentile coding-agent day in our Agent Burn Index, measured on OpenCode, Cline, Kilo and similar agents, September 13 to October 12, 2026. Priced at the list rates above; your own days will differ.
What else to expect
- Codex's shell and edit tools reach GLM through the Responses API.
- Web search is off on this connection, and the launcher offers one reasoning level, medium.
/modelinside a launcher session lists every served model, so GLM 5.3 Flash or a GPT model is one command away.
The catch
Z.ai sells its own GLM Coding Plan from $18 a month, but we found no Codex guide or Responses endpoint in Z.ai's docs on October 14, 2026. If you want GLM in an agent Z.ai supports, Claude Code on Z.ai's own endpoint is the direct route.
Where Standard Compute fits
Standard Compute serves GLM 5.3 and its Flash sibling over the Responses API Codex needs, next to GPT and Claude on one key.
FAQ
Can I use wire_api = "chat" with Z.ai's OpenAI-compatible endpoint?
Not on current Codex releases, which only use the Responses API for custom providers. That is why a Chat Completions endpoint alone no longer works.
Does GLM 5.3 work with Codex's apply_patch edits?
Codex sends its edit and shell tools to the model as tool definitions, and GLM 5.3 answers with tool calls. How cleanly any non-GPT model follows Codex's patch format varies, so try it on a real task first.