Nemotron 3 Ultra
Nemotron 3 Ultra is a language model with 550B parameters from NVIDIA. It is currently free at 3 providers with up to 1M tokens of context. The model supports step-by-step reasoning, is suited to coding, supports tool calling. Its weights are open, so it can also run locally.
- Developer
- NVIDIA
- Parameters
- 550B / 55B active
- Context
- 1M tokens
- Type
- Text
- License
- Open weights
Where to use Nemotron 3 Ultra for free
Each row is a separate free endpoint with its own limits.
| Provider | Context | Limit | Card | |
|---|---|---|---|---|
OpenRouter nvidia/nemotron-3-ultra-550b-a55b:free |
1M | Free tier | No | Open |
NVIDIA Build nvidia/nemotron-3-ultra-550b-a55b |
1M | Free serverless inference for development; limits and model availability may change. | No | Open |
UnoRouter nemotron-3-ultra-550b-a55b:free |
1M | No-card free models use shared capacity and per-model limits; availability can change. | No | Open |
Quick start
curl https://openrouter.ai/api/v1/chat/completions \
-H "Authorization: Bearer $OPENROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "nvidia/nemotron-3-ultra-550b-a55b:free",
"messages": [{"role": "user", "content": "Hello!"}]
}'from openai import OpenAI
client = OpenAI(
base_url="https://openrouter.ai/api/v1",
api_key="YOUR_OPENROUTER_API_KEY",
)
reply = client.chat.completions.create(
model="nvidia/nemotron-3-ultra-550b-a55b:free",
messages=[{"role": "user", "content": "Hello!"}],
)
print(reply.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://openrouter.ai/api/v1",
apiKey: process.env.OPENROUTER_API_KEY,
});
const reply = await client.chat.completions.create({
model: "nvidia/nemotron-3-ultra-550b-a55b:free",
messages: [{ role: "user", content: "Hello!" }],
});
console.log(reply.choices[0].message.content);Example for OpenRouter. Any OpenAI-compatible client works: Cursor, Cline, OpenCode, Open WebUI.
Benchmarks
Source: Artificial Analysis · Nemotron 3 Ultra 550B A55B (Reasoning)
Similar free models
FAQ
Is Nemotron 3 Ultra free?
Yes. As of September 26, 2026, Nemotron 3 Ultra is free at OpenRouter, NVIDIA Build, UnoRouter. Each provider has its own limits — see the table above.
What is the context window of Nemotron 3 Ultra?
The largest context among free routes is 1,048,576 tokens. Some providers cap it lower.
How do I call Nemotron 3 Ultra via API?
Use any OpenAI-compatible client with base URL https://openrouter.ai/api/v1, model ID nvidia/nemotron-3-ultra-550b-a55b:free, and a OpenRouter key.
How good is Nemotron 3 Ultra?
Artificial Analysis Intelligence Index: 22.9, GPQA Diamond: 87%. For comparison, the strongest free model in the catalog right now is GLM 5.3.
Is Nemotron 3 Ultra open-weight?
Yes, the weights are published — you can download and run it locally, for example with Ollama or LM Studio.