NVIDIA NIM API
first_party
Getting credentials
api-key
api_key via header (header Authorization: Bearer)
Key prefix: nvapi-
Environment: NVIDIA_API_KEY
Generate nvapi- keys at build.nvidia.com (free with the NVIDIA Developer Program).
Endpoints
| ID | URL | Protocol | Auth |
|---|---|---|---|
| v1-chat-completions | https://integrate.api.nvidia.com/v1/chat/completions | openai-chat | any |
API surfaces
textstreamingembeddings
Quirks
- No pay-as-you-go: hosted access is free-tier rate limited (~40 RPM soft); production requires deploying NIM yourself or NVIDIA AI Enterprise. (docs)
- The real reasoning switch is chat_template_kwargs (extra_body): enable_thinking (Qwen-style), medium_effort (Nemotron-3), reasoning_strength (muse-glimmer). reasoning_effort is accepted by the schema but silently ignored on some models. (docs)
- Hosted temperature range is 0-1 (default 0.95), not 0-2. Old catalog entries (e.g. meta/llama2-70b) still listed but 404. (docs)
Models offered
| Model | Wire ID | Endpoint | Status | Reasoning |
|---|---|---|---|---|
| deepseek/deepseek-v4-pro | deepseek-ai/deepseek-v4-pro | v1-chat-completions | retired | adaptive · default on · returns reasoning_content · round-trips reasoning_content |
| google/gemma-4 | google/gemma-4-31b-it | v1-chat-completions | ga | adaptive · default on · returns reasoning_content |
| meta/llama-3-3-70b | meta/llama-3.3-70b-instruct | v1-chat-completions | retired | none · default off · returns hidden |
| meta/muse-glimmer-30b | meta/muse-glimmer-30b | v1-chat-completions | ga | effort · default on · returns reasoning_content |
| minimax/minimax-m3 | minimaxai/minimax-m3 | v1-chat-completions | retired | effort · default on · mandatory · returns reasoning_content |
| moonshot/kimi-k2-6 | moonshotai/kimi-k2.6 | v1-chat-completions | ga | toggle · default on · returns reasoning_content |
| moonshot/kimi-k3 | moonshotai/kimi-k3 | v1-chat-completions | ga | toggle · default on · returns reasoning_content |
| nvidia/nemotron-3-5-lightning | nvidia/nemotron-3.5-lightning-30b-a3b | v1-chat-completions | ga | toggle · default on · returns reasoning_content |
| nvidia/nemotron-3-super | nvidia/nemotron-3-super-120b-a12b | v1-chat-completions | ga | toggle · default on · returns reasoning_content |
| nvidia/nemotron-3-ultra | nvidia/nemotron-3-ultra-550b-a55b | v1-chat-completions | ga | toggle · default on · returns reasoning_content |
| openai/gpt-oss-120b | openai/gpt-oss-120b | v1-chat-completions | retired | toggle · default on · returns reasoning_content |
| openai/gpt-oss-20b | openai/gpt-oss-20b | v1-chat-completions | ga | toggle · default on · returns reasoning_content |
| poolside/laguna-xs-2-1 | poolside/laguna-xs-2.1 | v1-chat-completions | ga | adaptive · default on · returns hidden |
| stepfun/step-3-7-flash | stepfun-ai/step-3.7-flash | v1-chat-completions | retired | effort · default on · mandatory · returns reasoning_content |