AI TOOLS RADAR
Explore
All models / NVIDIA
nvidia logo
nvidia

NVIDIA: Nemotron 3 Ultra

NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). $0.50 per million input tokens, $2.20 per million output tokens. 262,144 token context window, maximum output of 16,384 tokens. Higher uptime with 5 providers. Includes independent benchmarks from Artificial Analysis.

View source on OpenRouter Hugging Face
Overview

Model specifications

Context
262,144 tokens
Maximum output
16,384 tokens
Architecture
text->text
Tokenizer
Other
Knowledge cutoff
Not provided
Moderated
No
OPENROUTER

Complete pricing

Synchronized OpenRouter rates. Token prices are shown per one million tokens.

Input
$0.5
/M tokens
Output
$2.2
/M tokens
Cache read
$0.1
/M tokens
API

Model configuration

Default parameters

temperature
1
top_p
0.95

Reasoning

Moderated
No
Default parameters
high
Capabilities
high, medium
API

Quick start

Call this exact model ID through OpenRouter’s OpenAI-compatible API.

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","messages":[{"role":"user","content":"Hello!"}]}'
Capabilities

Input → Output

Input
text
Output
text
Reasoning

Yes · high · medium

API

Supported API parameters

frequency_penaltyinclude_reasoninglogit_biasmax_tokensmin_ppresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_p
Available providers

3 providers

Live providers on OpenRouter

Provider availability, latency, throughput and routing can change continuously. Open the source page for current operational data.

OpenRouter

DeepInfra

fp4
Context
262K
Maximum output
16K
Input
$0.500
Output
$2.20
Cache read
$0.100
Cache write
Not provided

BaseTen

fp4
Context
203K
Maximum output
183K
Input
$0.600
Output
$2.40
Cache read
$0.120
Cache write
Not provided

Venice

fp8
Context
256K
Maximum output
33K
Input
$0.625
Output
$3.13
Cache read
$0.188
Cache write
Not provided
nvidia

models

All models
nvidia logonvidia

NVIDIA: Nemotron 3.5 ASR Streaming Multilingual 0.6B

Nemotron 3.5 ASR Streaming Multilingual 0.6B is a speech recognition model from NVIDIA. Its prompt-conditioned, cache-aware FastConformer-RNNT design targets low-latency transcription across more than 40 languages for real-time captioning, voice...

Context
Not provided
Input
audio
Output
transcription
Input: $3.33Output: Freeper 1M tokens
View model details →
nvidia logonvidia

NVIDIA: Nemotron 3.5 Lightning

NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that...

Context
262K
Input
text
Output
text
Input: $0.080Output: $0.200per 1M tokens
View model details →
nvidia logonvidia

NVIDIA: Nemotron 3.5 Lightning (free)

NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that...

Context
1M
Input
text
Output
text
Input: FreeOutput: Freeper 1M tokens
View model details →
nvidia logonvidia

NVIDIA: Nemotron 3 Embed 1B (free)

NVIDIA Nemotron 3 Embed 1B is an open text embedding model from NVIDIA, optimized for high-throughput, low-latency retrieval. It is suited for enterprise search, RAG, code retrieval, and agentic retrieval...

Context
33K
Input
text
Output
embeddings
Input: FreeOutput: Freeper 1M tokens
View model details →