Point Zed at Tium
You need the base URL https://api.tium.ai/v1 and an API key from your dashboard. Keys are shown once at creation.
1.Add the provider to settings.json
Open the command palette and run zed: open settings, then add the block above. Any slug from the catalog can go in available_models; add one object per model you want offered.
settings.json{ "language_models": { "openai_compatible": { "tium": { "api_url": "https://api.tium.ai/v1", "available_models": [ { "name": "glm-5.3-flash", "display_name": "GLM-5.3 Flash", "max_tokens": 1000000 } ] } } } }2.Enter the key in the settings UI, not the file
Zed's own documentation is explicit that API keys do not belong in settings.json. Open the agent panel's provider settings and paste the key there, or set the environment variable Zed generates for the provider.
What Zed gets wrong
Tool-specific behaviour that surprises people, and what to do about it.
max_tokens here is the context window
Despite the name it is the input ceiling Zed sizes conversations against, not the output limit. Setting it to our served context is what stops Zed from packing a request we will refuse.