Ollama Cloud
first_party
Getting credentials
api-key
api_key via header (header Authorization: Bearer)
Environment: OLLAMA_API_KEY
Create keys at ollama.com/settings/keys (they do not currently expire).
Endpoints
| ID | URL | Protocol | Auth |
|---|---|---|---|
| v1-chat-completions | https://ollama.com/v1/chat/completions | openai-chat | any |
API surfaces
textstreaming
Plan
- Price
- $20 / monthly
- Quota
- Free: 1 concurrent model; Pro $20: 3; Max $100: 10. Sessions reset every 5 hours plus weekly limits; usage weighted by model cost level; extra balance purchasable on Pro/Max.
- Notes
- Zero data retention; no training on prompts; US-primary hosting.
- Docs
- https://ollama.com/cloud
Quirks
- Native API at https://ollama.com/api (chat/generate/tags) uses the think parameter (low/medium/high/max); gpt-oss models accept only low/medium/high and cannot disable thinking. An Anthropic-compatible API is also documented. https://ollama.com/api/v1 does NOT exist (404). (docs)
- The OpenAI-compatible /v1 accepts reasoning_effort (high/medium/low/max/none) and reasoning.effort. Reasoning returns in message.thinking on the native API — not reasoning_content. Cloud embeddings are unverified: /api/embed may not be authorized for cloud keys. (docs)
- Usage levels (Low, Medium, High, Extra High) weight billing — no per-Mtok pricing. Full catalog at ollama.com/search?c=cloud (also gemma4, nemotron-3 family, mistral-large-3 — the only non-thinking cloud model). (docs)
- Model IDs dropped the -cloud suffix (verified 2026-08-19); DeepSeek and Qwen entries carry snapshot/variant suffixes (:0731/:0813/:397b/:preview) and full-catalog drift is visible in the daily sync. (docs)
Models offered
| Model | Wire ID | Endpoint | Status | Reasoning |
|---|---|---|---|---|
| deepseek/deepseek-v4-flash | deepseek-v4-flash:0731 | v1-chat-completions | ga | effort · default on · returns hidden |
| deepseek/deepseek-v4-pro | deepseek-v4-pro:0813 | v1-chat-completions | ga | effort · default on · returns hidden |
| google/gemma-4 | gemma4:31b | v1-chat-completions | ga | effort · default on · returns hidden |
| minimax/minimax-m2-7 | minimax-m2.7 | v1-chat-completions | ga | effort · default on · returns hidden |
| minimax/minimax-m3 | minimax-m3 | v1-chat-completions | ga | effort · default on · returns hidden |
| mistral/mistral-large-3 | mistral-large-3:675b | v1-chat-completions | ga | none · default off · returns hidden |
| moonshot/kimi-k2-6 | kimi-k2.6 | v1-chat-completions | ga | effort · default on · returns hidden |
| moonshot/kimi-k2-7-code | kimi-k2.7-code | v1-chat-completions | ga | effort · default on · returns hidden |
| moonshot/kimi-k3 | kimi-k3 | v1-chat-completions | ga | effort · default on · returns hidden |
| nvidia/nemotron-3-super | nemotron-3-super | v1-chat-completions | ga | effort · default on · returns hidden |
| nvidia/nemotron-3-ultra | nemotron-3-ultra | v1-chat-completions | ga | effort · default on · returns hidden |
| openai/gpt-oss-120b | gpt-oss:120b | v1-chat-completions | ga | effort · default on · mandatory · returns hidden |
| openai/gpt-oss-20b | gpt-oss:20b | v1-chat-completions | ga | effort · default on · mandatory · returns hidden |
| zai/glm-5-1 | glm-5.1 | v1-chat-completions | ga | effort · default on · returns hidden |
| zai/glm-5-2 | glm-5.2 | v1-chat-completions | ga | effort · default on · returns hidden |
| zai/glm-5-3 | glm-5.3 | v1-chat-completions | ga | effort · default on · mandatory · returns hidden |
| zai/glm-5-3-flash | glm-5.3-flash | v1-chat-completions | ga | effort · default on · mandatory · returns hidden |