NVIDIA NIM 免费 API 详情
实测于 2026-08-27 00:35 +08:00（来源：oneapi 生产 DB Channel 表直读 + 官方 API https://integrate.api.nvidia.com/v1/models 无鉴权实测 83）
渠道 slug：nvidia-nim | 详情页 URL：/free-llm-api-rankings/nvidia-nim | Base URL：https://integrate.api.nvidia.com/v1
1. 渠道介绍
NVIDIA NIM（NVIDIA Inference Microservice）通过 integrate.api.nvidia.com/v1 提供 83 个免费模型——2026-08-27 00:35 无鉴权实测（83 total，33 个 nvidia/ 官方）。免费层覆盖对话 57、嵌入 7、审核/安全 5、翻译 3、视觉 6，含 Llama、Gemma、Qwen、Mistral、DeepSeek、Kimi、Phi 及完整 Nemotron 家族。60 RPM，兼容 OpenAI 的 /v1/chat/completions 与 /v1/models。
2. 实测支持模型列表
口径：models_total = 83（Channel.models 全量直读 + 官方 /v1/models 83 交叉验证）；chat = 57；embedding = 7；moderation = 5；translation = 3；vision = 6
对话模型（57 个）
- 01-ai/yi-large
- ai21labs/jamba-1.5-large-instruct
- aisingapore/sea-lion-7b-instruct
- bigcode/starcoder2-15b
- databricks/dbrx-instruct
- deepseek-ai/deepseek-coder-6.7b-instruct
- deepseek-ai/deepseek-v4-flash-0731
- deepseek-ai/deepseek-v4-pro-0813
- google/codegemma-1.1-7b
- google/codegemma-7b
- google/diffusiongemma-26b-a4b-it
- google/gemma-2b
- google/gemma-3-12b-it
- google/gemma-3-4b-it
- google/gemma-4-31b-it
- google/recurrentgemma-2b
- ibm/granite-3.0-3b-a800m-instruct
- ibm/granite-3.0-8b-instruct
- ibm/granite-34b-code-instruct
- ibm/granite-8b-code-instruct
- meta/codellama-70b
- meta/llama2-70b
- meta/muse-glimmer-30b
- microsoft/phi-3.5-moe-instruct
- minimaxai/minimax-m3
- mistralai/codestral-22b-instruct-v0.1
- mistralai/mistral-7b-instruct-v0.3
- mistralai/mistral-large
- mistralai/mistral-large-2-instruct
- mistralai/mistral-nemotron
- mistralai/mixtral-8x22b-v0.1
- moonshotai/kimi-k2.6
- moonshotai/kimi-k3
- nv-mistralai/mistral-nemo-12b-instruct
- nvidia/cosmos-reason2-8b
- nvidia/ising-calibration-1.5-31b
- nvidia/llama-3.1-nemotron-51b-instruct
- nvidia/llama-3.1-nemotron-70b-instruct
- nvidia/llama-3.1-nemotron-ultra-253b-v1
- nvidia/llama3-chatqa-1.5-70b
- nvidia/mistral-nemo-minitron-8b-8k-instruct
- nvidia/nemotron-3-nano-30b-a3b
- nvidia/nemotron-3-nano-omni-30b-a3b-reasoning
- nvidia/nemotron-3-super-120b-a12b
- nvidia/nemotron-3-ultra-550b-a55b
- nvidia/nemotron-3.5-lightning-30b-a3b
- nvidia/nemotron-4-340b-instruct
- nvidia/nemotron-4-340b-reward
- nvidia/nemotron-nano-3-30b-a3b
- openai/gpt-oss-120b
- openai/gpt-oss-20b
- poolside/laguna-xs-2.1
- writer/palmyra-creative-122b
- writer/palmyra-fin-70b-32k
- writer/palmyra-med-70b
- writer/palmyra-med-70b-32k
- zyphra/zamba2-7b-instruct
嵌入模型（7 个）
- nvidia/embed-qa-4
- nvidia/llama-3.2-nemoretriever-1b-vlm-embed-v1
- nvidia/llama-3.2-nv-embedqa-1b-v1
- nvidia/llama-nemotron-embed-vl-1b-v2
- nvidia/nemotron-3-embed-1b
- nvidia/nv-embedqa-mistral-7b-v2
- snowflake/arctic-embed-l
审核/安全模型（5 个）
- meta/llama-guard-4-12b
- nvidia/llama-3.1-nemoguard-8b-content-safety
- nvidia/llama-3.1-nemoguard-8b-topic-control
- nvidia/llama-3.1-nemotron-safety-guard-8b-v3
- nvidia/nemotron-3.5-content-safety
翻译模型（3 个）
- nvidia/riva-translate-4b-instruct
- nvidia/riva-translate-4b-instruct-v1.1
- nvidia/riva-translate-4b-instruct-v2
视觉模型（6 个）
- meta/llama-3.2-11b-vision-instruct
- meta/llama-3.2-90b-vision-instruct
- microsoft/phi-3-vision-128k-instruct
- nvidia/cosmos-reason2-8b
- nvidia/neva-22b
- nvidia/vila
3. 注册方法
- 访问 https://build.nvidia.com 使用 NVIDIA 账号登录。
- 在 https://integrate.api.nvidia.com 生成 API Key（请求头
Authorization: Bearer nvapi-...）。
- 免费层无需信用卡，限速 60 RPM。
4. 使用方法
Base URL: https://integrate.api.nvidia.com/v1
兼容 OpenAI。模型列表：GET /v1/models（无需鉴权，返回 83）。对话：POST /v1/chat/completions 携带 model 与 messages。示例：model: "openai/gpt-oss-120b" 或 "nvidia/nemotron-3-nano-30b-a3b"。
5. 实测数据
实测于 2026-08-27 00:35 +08:00，官方 /v1/models 无鉴权 200，返回 83。
| 指标 |
值 |
备注 |
| HTTP 状态码 |
200 |
GET /v1/models 无鉴权 |
| 模型总数 |
83 |
含 33 个 nvidia/ 官方 |
| 对话模型 |
57 |
含 Llama/Qwen/Mistral/DeepSeek 等 |
| 首 token 延迟 |
待填 |
需 /chat/completions 实测 |
| RPM 限制 |
60 |
Channel 表 maxRpm |
| 免费层 |
✅ |
无需信用卡，build.nvidia.com |
6. 优缺点总结
优点：
- 83 个模型，免费池最大（57 对话 + 6 视觉 + 7 嵌入）
- 33 个 nvidia/ 官方 Nemotron 模型
- OpenAI 兼容，/v1/models 无需鉴权
- NIM GPU 加速
缺点：
- 部分模型已 EOL（如 meta/llama-3.1-70b-instruct 于 2026-08-26 410）
- 翻译/视觉为专用模型，非通用对话
本页由 IT 于 2026-08-27 00:35 四重复验生成，数据以 oneapi 生产 DB Channel 表直读 + 官方 API 实测为准。
NVIDIA NIM Free API Details
Verified on 2026-08-27 00:35 +08:00 (source: OneAPI production DB Channel table direct read + official API https://integrate.api.nvidia.com/v1/models no-auth verified 83)
Channel slug:nvidia-nim | Detail URL:/free-llm-api-rankings/nvidia-nim | Base URL:https://integrate.api.nvidia.com/v1
1. Channel Overview
NVIDIA NIM (NVIDIA Inference Microservice) exposes 83 free models via integrate.api.nvidia.com/v1 — verified 2026-08-27 00:35 without auth (83 total, 33 nvidia/ official). The free tier covers chat (57), embedding (7), moderation/safety (5), translation (3) and vision (6), including Llama, Gemma, Qwen, Mistral, DeepSeek, Kimi, Phi and the full Nemotron family. 60 RPM, OpenAI-compatible /v1/chat/completions and /v1/models.
2. Verified Supported Models
Scope: models_total = 83 (full Channel.models direct read + official /v1/models 83 cross-verified); chat = 57; embedding = 7; moderation = 5; translation = 3; vision = 6
Chat Models (57)
- 01-ai/yi-large
- ai21labs/jamba-1.5-large-instruct
- aisingapore/sea-lion-7b-instruct
- bigcode/starcoder2-15b
- databricks/dbrx-instruct
- deepseek-ai/deepseek-coder-6.7b-instruct
- deepseek-ai/deepseek-v4-flash-0731
- deepseek-ai/deepseek-v4-pro-0813
- google/codegemma-1.1-7b
- google/codegemma-7b
- google/diffusiongemma-26b-a4b-it
- google/gemma-2b
- google/gemma-3-12b-it
- google/gemma-3-4b-it
- google/gemma-4-31b-it
- google/recurrentgemma-2b
- ibm/granite-3.0-3b-a800m-instruct
- ibm/granite-3.0-8b-instruct
- ibm/granite-34b-code-instruct
- ibm/granite-8b-code-instruct
- meta/codellama-70b
- meta/llama2-70b
- meta/muse-glimmer-30b
- microsoft/phi-3.5-moe-instruct
- minimaxai/minimax-m3
- mistralai/codestral-22b-instruct-v0.1
- mistralai/mistral-7b-instruct-v0.3
- mistralai/mistral-large
- mistralai/mistral-large-2-instruct
- mistralai/mistral-nemotron
- mistralai/mixtral-8x22b-v0.1
- moonshotai/kimi-k2.6
- moonshotai/kimi-k3
- nv-mistralai/mistral-nemo-12b-instruct
- nvidia/cosmos-reason2-8b
- nvidia/ising-calibration-1.5-31b
- nvidia/llama-3.1-nemotron-51b-instruct
- nvidia/llama-3.1-nemotron-70b-instruct
- nvidia/llama-3.1-nemotron-ultra-253b-v1
- nvidia/llama3-chatqa-1.5-70b
- nvidia/mistral-nemo-minitron-8b-8k-instruct
- nvidia/nemotron-3-nano-30b-a3b
- nvidia/nemotron-3-nano-omni-30b-a3b-reasoning
- nvidia/nemotron-3-super-120b-a12b
- nvidia/nemotron-3-ultra-550b-a55b
- nvidia/nemotron-3.5-lightning-30b-a3b
- nvidia/nemotron-4-340b-instruct
- nvidia/nemotron-4-340b-reward
- nvidia/nemotron-nano-3-30b-a3b
- openai/gpt-oss-120b
- openai/gpt-oss-20b
- poolside/laguna-xs-2.1
- writer/palmyra-creative-122b
- writer/palmyra-fin-70b-32k
- writer/palmyra-med-70b
- writer/palmyra-med-70b-32k
- zyphra/zamba2-7b-instruct
Embedding Models (7)
- nvidia/embed-qa-4
- nvidia/llama-3.2-nemoretriever-1b-vlm-embed-v1
- nvidia/llama-3.2-nv-embedqa-1b-v1
- nvidia/llama-nemotron-embed-vl-1b-v2
- nvidia/nemotron-3-embed-1b
- nvidia/nv-embedqa-mistral-7b-v2
- snowflake/arctic-embed-l
Moderation/Safety Models (5)
- meta/llama-guard-4-12b
- nvidia/llama-3.1-nemoguard-8b-content-safety
- nvidia/llama-3.1-nemoguard-8b-topic-control
- nvidia/llama-3.1-nemotron-safety-guard-8b-v3
- nvidia/nemotron-3.5-content-safety
Translation Models (3)
- nvidia/riva-translate-4b-instruct
- nvidia/riva-translate-4b-instruct-v1.1
- nvidia/riva-translate-4b-instruct-v2
Vision Models (6)
- meta/llama-3.2-11b-vision-instruct
- meta/llama-3.2-90b-vision-instruct
- microsoft/phi-3-vision-128k-instruct
- nvidia/cosmos-reason2-8b
- nvidia/neva-22b
- nvidia/vila
3. Registration
- Visit https://build.nvidia.com and sign in with NVIDIA account.
- Generate API key at https://integrate.api.nvidia.com (header
Authorization: Bearer nvapi-...).
- Free tier is available without credit card; rate limit 60 RPM.
4. Usage
Base URL: https://integrate.api.nvidia.com/v1
OpenAI-compatible. List models: GET /v1/models (no auth required, returns 83). Chat: POST /v1/chat/completions with model and messages. Example: model: "openai/gpt-oss-120b" or "nvidia/nemotron-3-nano-30b-a3b".
5. Verified Metrics
实测于 2026-08-27 00:35 +08:00，官方 /v1/models 无鉴权 200，返回 83。
| Metric |
Value |
Notes |
| HTTP 状态码 |
200 |
GET /v1/models 无鉴权 |
| 模型总数 |
83 |
含 33 个 nvidia/ 官方 |
| 对话模型 |
57 |
含 Llama/Qwen/Mistral/DeepSeek 等 |
| 首 token 延迟 |
待填 |
需 /chat/completions 实测 |
| RPM 限制 |
60 |
Channel 表 maxRpm |
| 免费层 |
✅ |
无需信用卡，build.nvidia.com |
6. Pros & Cons
Pros:
- 83 models, largest free pool (57 chat + 6 vision + 7 embedding)
- 33 nvidia/ official Nemotron models
- OpenAI-compatible, no auth for /v1/models
- GPU-accelerated NIM
Cons:
- Some models EOL (e.g. meta/llama-3.1-70b-instruct 410 on 2026-08-26)
- Translation/vision models are specialized, not general chat
本页由 IT 于 2026-08-27 00:35 四重复验生成，数据以 oneapi 生产 DB Channel 表直读 + 官方 API 实测为准。