Skip to main content

Nemotron 3 Ultra 550B A55B

NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it supports text input and output with a context window of up to 1M tokens. It is suited for long-running agentic workflows, including agent orchestration, coding agents, deep research, and complex enterprise tasks. It is particularly strong at multi-step reasoning and planning, with high-throughput inference designed for high-volume agent pipelines. It is part of the NVIDIA Nemotron family of open models for agentic AI.

Input / output price
$0.50 / $2.20 per 1M
Context
1M tokens
Released
Jun 4, 2026
Agent setup

Paste into Claude Code, Cursor, or Codex

Help me set up the Respan gateway: fetch and follow https://www.respan.ai/docs/agent-setup.md. Use model "nemotron-3-ultra-550b-a55b".

Read docs

Providers

6 providers, cheapest first, retired then OpenRouter last, priced per 1M tokens after gateway discounts.

ProviderInputOutput
Concentrateconcentrate/nemotron-3-ultra-nvfp4
$0.50/M$2.20/M
DeepInfradeepinfra/nemotron-3-ultra-550b-a55b
$0.50/M$2.20/M
Basetenbaseten/nemotron-3-ultra-550b-a55b
$0.60/M$2.40/M
Nebiusnebius/nemotron-3-ultra-550b-a55b
$1.00/M$3.00/M
FireworksRetiredfireworks/nemotron-3-ultra-550b-a55b
$0.60/M$2.40/M

Is a provider down? Fireworks AI status, OpenRouter status

Pricing

Averaged across 6 providers, in USD per 1M tokens after gateway discounts.

Average input
$0.62 per 1M
Average output
$2.40 per 1M
Average cache read
$0.11 per 1M
Average cache write
$0.31 per 1M

Cost by volume

Total cost from 0 to 1T tokens on the cheapest and most expensive provider.

Highest$1.00/MNebiusLowest$0.50/MConcentrate, DeepInfra, and OpenRouter

Throughput and latency

30-day averages from requests through the Respan gateway. Under each figure, every model on the gateway by its average, this one in blue.

Output speed

Higher is better

24tokens/sacross providers

Faster than 53% of the other 595 models on the gateway

Response latency

Lower is better

6.2sacross providers

Faster than 26% of the other 756 models on the gateway

Benchmarks

The Respan Index plus 22 published scores from 8 sources, each linked to where it was published.

28.4/ 100

Respan Index#97 of 176

176 ranked models. The shaded band is this model's 95% interval, 23.4 to 35.0. Kimi K2.5 and Hy3 preview fall inside it, so the gap is within the margin of error.
Nemotron 3 Ultra 550B A55B by category: Coding 6.3, Agents 33.0, Reasoning 30.8, Knowledge & language 35.1, Human preference 49.4
  1. 97NVIDIA Nemotron 3 Ultra 550B-A55B NVFP428.4
  2. 96Kimi K2.529.3
  3. 98Hy3 preview28.3

Published scores

BenchmarkScore
LiveBench CodingLiveBench70.7%
SWE-bench Verifiedbuild.nvidia.com69.7%
Terminal-Bench 2.1build.nvidia.com53.9%
WeirdMLHåvard Tveit Ihle43.5%

Compare Nemotron 3 Ultra 550B A55B: NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 vs Claude Opus 4.8, NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 vs NVIDIA Nemotron 3.5 Lightning 30B-A3B NVFP4, NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 vs NVIDIA Nemotron 3.5 Super VL 120B-A12B NVFP4, NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 vs NVIDIA Nemotron 3 Super 120B-A12B, NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 vs NVIDIA Nemotron 3 Nano 30B-A3B BF16

Apps

Use Nemotron 3 Ultra 550B A55B in the tools you already run. Each guide points the app at the Respan gateway.

Setup guides for more apps are in the integration docs.

Nemotron 3 Ultra 550B A55B FAQ

Short answers, built from the data on this page.

Nemotron 3 Ultra 550B A55B is a model from NVIDIA, released on Jun 4, 2026. NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it supports text input and output with a context window of up to 1M tokens. It is suited for long-running agentic workflows, including agent orchestration, coding agents, deep research, and complex enterprise tasks. It is particularly strong at multi-step reasoning and planning, with high-throughput inference designed for high-volume agent pipelines. It is part of the NVIDIA Nemotron family of open models for agentic AI.

On the Respan gateway, Nemotron 3 Ultra 550B A55B costs from $0.50 per 1M input tokens and $2.20 per 1M output tokens on Concentrate, DeepInfra, and OpenRouter, after gateway discounts. Averaged across its 6 providers, it costs $0.62 per 1M input tokens and $2.40 per 1M output tokens.

Nemotron 3 Ultra 550B A55B is available through 6 providers: Concentrate, DeepInfra, Baseten, Nebius, Fireworks, and OpenRouter.

Nemotron 3 Ultra 550B A55B accepts up to 1M tokens of context on its largest provider route.

Point any OpenAI-compatible SDK at https://gateway.respan.ai/v1 with your Respan API key and set the model to "nemotron-3-ultra-550b-a55b". To set it up with a coding agent, copy the agent setup prompt at the top of this page.

Nemotron 3 Ultra 550B A55B scores 28.4 out of 100 on the Respan Index v1.1, ranking #97 of 176 models. By category: Coding 6.3, Agents 33.0, Reasoning 30.8, Knowledge & language 35.1, and Human preference 49.4. The Respan Index combines independent and lab-reported benchmark results into one score, using benchmark data collected by Respan.

Nemotron 3 Ultra 550B A55B has 22 published benchmark scores, including LiveBench Coding 70.7%, BrowseComp 41.4%, and GPQA Diamond 87.9%. Each score links to where it was published; the data is collected by Respan.

Certified where it counts.

Built to meet the security and privacy standards that enterprise and healthcare teams require.

ISO 27001

The internationally recognized standard for information security management.

SOC 2

Secure, compliant management of your data across all of our systems.

GDPR

Operated under GDPR, the world's strictest standard for data privacy.

HIPAA

HIPAA compliant, with a BAA available for healthcare teams.

Works with your entire stack

Use Respan with your favorite frameworks and tools.

AnthropicAI21 LabsxAIVercel AI SDK
LangChainOpenRouterAssemblyAICursor
LiteLLMGroqBasetenMem0
CohereCogneeLlamaIndexFireworks
DeepSeekLinkupHaystackTogether AI
PerplexityInference.netPostHogRubyLLM
SlackMastraAzure OpenAIMistral
AWS BedrockMoonshotHyperspellBAML
ModaInstructorGoogle GeminiParasail
ReductoZapierAgnoNebius
OpenAINovita AIReplicateSuperagent

Built for AI agents.
Break less.
Ship more.