Nemotron 3 Ultra 550B a55b:thinking API

NVIDIA's Nemotron 3 Ultra 550B a55b:thinking costs $0.5 per million input tokens and $2.5 per million output tokens from 1 supplier. Each request through the OpenAI-compatible API goes to the offer with the lowest estimated charge that meets your rules.

About Nemotron 3 Ultra 550B a55b:thinking

Nemotron 3 Ultra 550B a55b:thinking is a language model from NVIDIA. Its release date is 4 June 2026. Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy. Its weights are open. Its training data ends in September 2025. It takes text and gives text. The largest offer takes 1,000,000 input tokens. It scores 11.8 on the Artificial Analysis Intelligence Index, #50 of 64 scored models on this exchange. It has 1 supplier on this exchange now.

Description, release date and licence: models.dev (https://models.dev/)

Facts

Model name
nemotron-3-ultra-550b-a55b:thinking
Maker
NVIDIA
Context window
1M tokens
Input price
$0.5 per million tokens
Output price
$2.5 per million tokens
Suppliers
1 supplier
Regions
us

Providers

Scores

Scores: Artificial Analysis (https://artificialanalysis.ai/)

Request

curl https://router.inference.ai.exchange/v1/chat/completions \
  -H "Authorization: Bearer $LIQUID_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nemotron-3-ultra-550b-a55b:thinking",
    "messages": [{"role":"user","content":"Say hello."}]
  }'

Questions

What is Nemotron 3 Ultra 550B a55b:thinking?
Nemotron 3 Ultra 550B a55b:thinking is a language model from NVIDIA. Its release date is 4 June 2026. Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy. Its weights are open. Its training data ends in September 2025.
How much does Nemotron 3 Ultra 550B a55b:thinking cost?
Offers now ask $0.5 per million input tokens and $2.5 per million output tokens. A job pays for the tokens it uses. The exchange picks the cheapest eligible offer.
What is the context length of Nemotron 3 Ultra 550B a55b:thinking?
The largest offer takes 1,000,000 input tokens. A smaller offer can take less. The exchange skips an offer that cannot take your message.
What can Nemotron 3 Ultra 550B a55b:thinking do?
Its providers list Reasoning effort, JSON object, JSON schema, Frequency penalty, Logit bias, Log probabilities, Min-p sampling, Presence penalty, Repetition penalty, Seed, Stop sequences, Top-k sampling, Top-p sampling, Text, Tools.
What inputs and outputs does Nemotron 3 Ultra 550B a55b:thinking support?
Text → Text.
Which providers serve Nemotron 3 Ultra 550B a55b:thinking?
It has 1 supplier now, with 1 offer. The table above ranks them by the cost of one agent-mix job.
How do I call Nemotron 3 Ultra 550B a55b:thinking?
Send "nemotron-3-ultra-550b-a55b:thinking" as the model in an OpenAI or Anthropic request. The API section links to setup steps for each client.
When was Nemotron 3 Ultra 550B a55b:thinking released?
Its creator released it on 4 June 2026.
When was Nemotron 3 Ultra 550B a55b:thinking listed?
The exchange first listed it on 10/10/2026. This is not its release date.

Related pages