Nemotron 3 Ultra 550B A55B

NVIDIAMCP Tool CallReasoning

nvidia/nemotron-3-ultra-550b-a55bOpen Weights

Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy

Benchmark Pricing$0.60 / 1M in$2.40 / 1M out$0.12 cache read

Context: 1,000,000 tokens•Max Output: 128,000 tokens•Knowledge: 2025-09•Released: 2026-06-04

TECHNICAL SPECIFICATIONS

Context Window1,000,0001M tokens

Max Output128,000128K tokens

LicenseOpen SourceSelf-hostable

Input Modalities

Text

Function / Tool CallingSupported (MCP Ready)

Structured OutputsGuaranteed Schema

MODEL CONTEXT PROTOCOL (MCP) READINESS

✓ Nemotron 3 Ultra 550B A55B natively supports OpenAI / Anthropic format tool calling and is fully compatible with MCP servers and agent loops.

MCP Context Window Recommendations

With a context capacity of 1,000,000 tokens, this model can handle extremely large MCP resource manifests, complete database schemas, and multiple active server tool definitions simultaneously without degradation.

STANDARDIZED BENCHMARKS & EVALS

SWE-Bench Verifiedresolved

70.7

SWE-Bench Multilingualresolve rate

67.7

Terminal-Benchsuccess rate

56.4

GPQAaccuracy

Humanity's Last Examaccuracy

26.7

Humanity's Last Examaccuracy

37.4

LiveCodeBenchpass@1

MMLU-Proaccuracy

86.8

BrowseCompaccuracy

44.4

IFBenchaccuracy

81.7

GDPvalwins or ties

46.7

API CALL INVOCATION EXAMPLE

// Example: Calling Nemotron 3 Ultra 550B A55B via OpenAI SDK
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.NVIDIA_API_KEY,
});

const response = await client.chat.completions.create({
  model: "nvidia/nemotron-3-ultra-550b-a55b",
  messages: [{ role: "user", content: "Hello! Can you inspect MCP tools?" }],
});

console.log(response.choices[0].message.content);