Llama 3.1 Nemotron Ultra 253B

Description

Llama 3.1 Nemotron Ultra 253B is Nvidia's flagship reasoning model, created by converting Meta's Llama 3.1 405B from dense to MoE architecture. Features 253B total parameters with ~17B active per token, Dynamic Thinking with toggleable reasoning, tool calling, and structured outputs.

Providers

Nebius

View provider details and policies.

Capabilities

EU-Compliant API Access

Use Llama 3.1 Nemotron Ultra 253B with a simple API call. OpenAI-compatible endpoint, EU data residency guaranteed.

const response = await fetch("https://api.eurouter.ai/api/v1/chat/completions", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "Authorization": `Bearer ${process.env.EUROUTER_API_KEY}`,
  },
  body: JSON.stringify({
    model: "llama-3.1-nemotron-ultra-253b",
    messages: [
      { role: "user", content: "Hello!" }
    ],
  }),
});

const data = await response.json();
console.log(data.choices[0].message.content);