Specs & substance

Developed by: NVIDIA
Model family: Nemotron
Use case: Flagship
Modality: Text
Context window: 128K tokens
Architecture: MoE NVFP4
Version: 3 Ultra
License: NVIDIA Open Model License
Pricing: $0.50 in · $2.50 out · $0.10 cache read
Released: Jun 2026
Endpoint: parasail-nemotron-3-ultra-550b-nvfp4

NVIDIA’s flagship open model optimized for enterprise reasoning workloads.

Nemotron 3 Ultra 550B is part of the Nemotron family by NVIDIA, and sits in the flagship category. It supports a 128K tokens context window and is built on a MoE NVFP4 architecture.

Key strengths: reasoning, enterprise. Parasail serves Nemotron 3 Ultra 550B on a global fleet of current-gen GPUs behind a single OpenAI-compatible endpoint — with per-token pricing, no minimums, and dedicated capacity options when you need guaranteed throughput.

Integrate

Drop-in via the OpenAI SDK

Point any OpenAI-compatible client at Parasail and change the model name. That's it.

python parasail · Nemotron 3 Ultra 550B

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.parasail.io/v1"
)

response = client.chat.completions.create(
    model="parasail-nemotron-3-ultra-550b-nvfp4",
    messages=[
        {"role": "user", "content": "Hello, what can you do?"}
    ],
    stream=True,
    max_tokens=1000
)

for chunk in response:
    if chunk.choices and chunk.choices[0].delta.content is not None:
        print(chunk.choices[0].delta.content, end="", flush=True)

bash parasail · Nemotron 3 Ultra 550B

curl https://api.parasail.io/v1/chat/completions \
  -H "Authorization: Bearer $PARASAIL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "parasail-nemotron-3-ultra-550b-nvfp4",
    "messages": [{"role": "user", "content": "Hello, what can you do?"}],
    "stream": true,
    "max_tokens": 1000
  }'

More models