Ollama Cloud in Codex — GLM, Kimi, DeepSeek | Codex Router

Ollama Cloud in Codex means adding hosted Ollama models — GLM, Kimi, MiniMax, DeepSeek and more — to the Codex model picker through Codex Router. The provider shares model families with other vendors but bills through an ollama.com account under its own quota.

Ollama CloudAPI keyhosted

What Ollama Cloud adds

Ollama Cloud can serve the same open-weight models that other providers offer, but through one ollama.com credential and quota. Matching entries such as DeepSeek V4 Pro intentionally coexist with vendor-direct routes because credentials and billing differ: a deepseek/deepseek-v4-pro key is billed by DeepSeek, while ollama-cloud/deepseek-v4-pro spends your Ollama Cloud allowance.

Enable Ollama Cloud

Install Codex Router first (see the installation guide), then store the account key and enable the provider:

1. Create an API key in your ollama.com account.
2. Store it with the hidden prompt.
3. Enable the provider and restart Codex.

./bin/model-router codex provider-key ollama-cloud set
./bin/model-router codex providers enable ollama-cloud

The base URL is https://ollama.com/v1. On Windows use ./model-router.ps1 codex with the same commands.

Registry models

These models ship in the registry and appear in the picker once the provider is enabled:

The router reads context windows from Ollama’s published model metadata rather than guessing them.

Candidate-route caution

GLM-5.3 and GLM-5.3-Flash on Ollama Cloud are candidate registry metadata. They still require a successful exact-route live proof — basic, streaming, forced-tool and stateless tool-result requests — before they should be described as certified. Treat them as unverified until that run passes.

Verify the integration

./bin/model-router codex doctor
./bin/model-router codex providers

The provider must report SHOW and ready. Then fully quit Codex, reopen it, create a new task, and check the picker. For a live, quota-consuming check:

./bin/test-model 'ollama-cloud/kimi-k3' --live --yes

If entries are missing, see models missing from the picker.

Add more models

Ollama Cloud ships a reviewed set, and other providers’ guides cover the same families on different billing:

• GLM in Codex — Z.ai coding, Z.ai API and Ollama Cloud routes
• DeepSeek in Codex — vendor-direct and cloud routes
• MiniMax in Codex — the MiniMax Token Plan route

Policy caution

ollama.com may restrict or block proxy traffic. Routing models through Codex Router is at your own risk; check Ollama’s terms before relying on it.

FAQ

What is Ollama Cloud in Codex? Ollama Cloud is an ollama.com account that serves hosted models through an OpenAI-compatible endpoint. Codex Router routes selected Ollama Cloud models into the Codex picker under the ollama-cloud provider.

How do I authenticate Ollama Cloud? Store an ollama.com account API key with ./bin/model-router codex provider-key ollama-cloud set, then enable the provider. The key bills through your ollama.com account.

Are the GLM-5.3 Ollama Cloud routes certified? Not yet. GLM-5.3 and GLM-5.3-Flash on Ollama Cloud are candidate registry metadata that still require exact-route live proof before being described as certified.

Related: models overview, DeepSeek in Codex, GLM in Codex, installation.