Skip to main content

Nvidia logoNvidia

Access 92 Nvidia models through Mastra's model router. Authentication is handled automatically using the NVIDIA_API_KEY environment variable.

Learn more in the Nvidia documentation.

.env
NVIDIA_API_KEY=your-api-key
src/mastra/agents/my-agent.ts
import { Agent } from "@mastra/core/agent";

const agent = new Agent({
  id: "my-agent",
  name: "My Agent",
  instructions: "You are a helpful assistant",
  model: "nvidia/abacusai/dracarys-llama-3_1-70b-instruct"
});

// Generate a response
const response = await agent.generate("Hello!");

// Stream a response
const stream = await agent.stream("Tell me a story");
for await (const chunk of stream) {
  console.log(chunk);
}
info

Mastra uses the OpenAI-compatible /chat/completions endpoint. Some provider-specific features may not be available. Check the Nvidia documentation for details.

Models
Direct link to Models

ModelContextToolsReasoningImageAudioVideoInput $/1MOutput $/1M
nvidia/abacusai/dracarys-llama-3_1-70b-instruct128K——
nvidia/baai/bge-m38K——
nvidia/black-forest-labs/flux_1-kontext-dev41K——
nvidia/black-forest-labs/flux_1-schnell77——
nvidia/black-forest-labs/flux_2-klein-4b41K——
nvidia/black-forest-labs/flux.1-dev4K——
nvidia/bytedance/seed-oss-36b-instruct262K——
nvidia/deepseek-ai/deepseek-v3.1-terminus128K——
nvidia/deepseek-ai/deepseek-v3.2164K——
nvidia/deepseek-ai/deepseek-v4-flash1.0M$0.14$0.28
nvidia/deepseek-ai/deepseek-v4-pro1.0M$2$3
nvidia/google/gemma-2-2b-it128K——
nvidia/google/gemma-3-27b-it131K——
nvidia/google/gemma-3n-e2b-it128K——
nvidia/google/gemma-3n-e4b-it128K——
nvidia/google/gemma-4-31b-it256K——
nvidia/google/google-paligemma128K——
nvidia/meta/esm2-650m128K——
nvidia/meta/esmfold128K——
nvidia/meta/llama-3.1-70b-instruct128K——
nvidia/meta/llama-3.1-8b-instruct16K——
nvidia/meta/llama-3.2-11b-vision-instruct128K——
nvidia/meta/llama-3.2-1b-instruct128K——
nvidia/meta/llama-3.2-3b-instruct33K——
nvidia/meta/llama-3.2-90b-vision-instruct128K——
nvidia/meta/llama-3.3-70b-instruct128K——
nvidia/meta/llama-4-maverick-17b-128e-instruct128K——
nvidia/meta/llama-guard-4-12b128K——
nvidia/microsoft/phi-4-mini-instruct131K——
nvidia/microsoft/phi-4-multimodal-instruct128K——
nvidia/minimaxai/minimax-m2.5205K——
nvidia/minimaxai/minimax-m2.7205K——
nvidia/mistralai/devstral-2-123b-instruct-2512262K——
nvidia/mistralai/magistral-small-250633K——
nvidia/mistralai/mistral-7b-instruct-v0366K——
nvidia/mistralai/mistral-large-3-675b-instruct-2512262K——
nvidia/mistralai/mistral-medium-3-instruct131K——
nvidia/mistralai/mistral-nemotron128K——
nvidia/mistralai/mistral-small-4-119b-2603128K——
nvidia/mistralai/mixtral-8x22b-instruct66K——
nvidia/mistralai/mixtral-8x7b-instruct33K——
nvidia/moonshotai/kimi-k2-instruct128K——
nvidia/moonshotai/kimi-k2-instruct-0905262K——
nvidia/moonshotai/kimi-k2-thinking262K——
nvidia/moonshotai/kimi-k2.6262K——
nvidia/nvidia/active-speaker-detection———
nvidia/nvidia/bevformer128K——
nvidia/nvidia/cosmos-predict1-5b———
nvidia/nvidia/cosmos-transfer1-7b———
nvidia/nvidia/cosmos-transfer2_5-2b———
nvidia/nvidia/gliner-pii128K——
nvidia/nvidia/llama-3_1-nemotron-safety-guard-8b-v3128K——
nvidia/nvidia/llama-3_2-nemoretriever-300m-embed-v133K——
nvidia/nvidia/llama-3_3-nemotron-super-49b-v1131K——
nvidia/nvidia/llama-3_3-nemotron-super-49b-v1_5131K——
nvidia/nvidia/llama-nemotron-embed-vl-1b-v233K——
nvidia/nvidia/llama-nemotron-rerank-vl-1b-v2128K——
nvidia/nvidia/magpie-tts-zeroshot———
nvidia/nvidia/nemotron-3-content-safety128K——
nvidia/nvidia/nemotron-3-nano-30b-a3b131K——
nvidia/nvidia/nemotron-3-nano-omni-30b-a3b-reasoning256K——
nvidia/nvidia/nemotron-3-super-120b-a12b262K$0.20$0.80
nvidia/nvidia/nemotron-content-safety-reasoning-4b128K——
nvidia/nvidia/nemotron-mini-4b-instruct128K——
nvidia/nvidia/nemotron-voicechat128K——
nvidia/nvidia/nv-embed-v133K——
nvidia/nvidia/nv-embedcode-7b-v133K——
nvidia/nvidia/nvidia-nemotron-nano-9b-v2131K——
nvidia/nvidia/rerank-qa-mistral-4b128K——
nvidia/nvidia/riva-translate-4b-instruct-v1_1128K——
nvidia/nvidia/sparsedrive128K——
nvidia/nvidia/streampetr128K——
nvidia/nvidia/studiovoice128K——
nvidia/nvidia/synthetic-video-detector———
nvidia/nvidia/usdcode128K——
nvidia/nvidia/usdvalidate———
nvidia/openai/gpt-oss-120b128K——
nvidia/openai/gpt-oss-20b131K——
nvidia/openai/whisper-large-v3———
nvidia/qwen/qwen-image———
nvidia/qwen/qwen-image-edit———
nvidia/qwen/qwen2.5-coder-32b-instruct128K——
nvidia/qwen/qwen3-coder-480b-a35b-instruct262K——
nvidia/qwen/qwen3-next-80b-a3b-instruct262K——
nvidia/qwen/qwen3-next-80b-a3b-thinking262K——
nvidia/qwen/qwen3.5-122b-a10b262K——
nvidia/qwen/qwen3.5-397b-a17b262K——
nvidia/sarvamai/sarvam-m128K——
nvidia/stepfun-ai/step-3.5-flash256K——
nvidia/upstage/solar-10_7b-instruct128K——
nvidia/z-ai/glm-5.1131K——
nvidia/z-ai/glm4.7205K——
92 available models

Advanced configuration
Direct link to Advanced configuration

Custom headers
Direct link to Custom headers

src/mastra/agents/my-agent.ts
const agent = new Agent({
  id: "custom-agent",
  name: "custom-agent",
  model: {
    url: "https://integrate.api.nvidia.com/v1",
    id: "nvidia/abacusai/dracarys-llama-3_1-70b-instruct",
    apiKey: process.env.NVIDIA_API_KEY,
    headers: {
      "X-Custom-Header": "value"
    }
  }
});

Dynamic model selection
Direct link to Dynamic model selection

src/mastra/agents/my-agent.ts
const agent = new Agent({
  id: "dynamic-agent",
  name: "Dynamic Agent",
  model: ({ requestContext }) => {
    const useAdvanced = requestContext.task === "complex";
    return useAdvanced
      ? "nvidia/z-ai/glm4.7"
      : "nvidia/abacusai/dracarys-llama-3_1-70b-instruct";
  }
});
On this page