Dedicated model · Available as managed deployment
NVIDIA's Nemotron 3 Ultra, a 550B mixture-of-experts with 55B active parameters, built for complex logic, coding and enterprise agent workflows, shipped in NVFP4 and BF16. Available as a managed deployment: validated on AxForge hardware and deployed on a multi-GPU system for your traffic only, as an OpenAI-compatible endpoint on hardware only you use, hosted in the EU.
Why Nemotron 3 Ultra
| Enterprise-grade reasoning | NVIDIA's largest open model, trained for complex multi-step logic and coding in corporate agent systems. |
|---|---|
| NVFP4-ready | Released in NVFP4 for NVIDIA's latest hardware, alongside a BF16 build. The NVFP4 build is efficient on the accelerators it was designed for. |
| NVIDIA open model licence | Ships under NVIDIA's open model licence; AxForge reviews it with you and deploys under it where it applies. |
Specifications
| Model | Nemotron 3 Ultra by nvidia |
|---|---|
| Modalities | Text |
| Sizes | 302.8B, 560.5B |
| Released | 2026-06 |
| Popularity | 404k downloads / month on HuggingFace |
| Licence | Commercial licence needed (the model's own licence); AxForge handles the licence where one is required |
| Availability | Available as managed deployment, operated by AxForge on hardware reserved for you |
| Hardware | Multi-GPU system (H100, H200 or B200 class) sourced against your request |
| Region | European region, placement confirmed with your request |
| Pricing | Hardware and managed service Quoted per deployment. Request deployment and we confirm both in writing. |
We publish only numbers we measure ourselves, and we haven't benchmarked this model on our nodes yet. For quality benchmarks, see the model card in the catalogue and the official card on HuggingFace.
Deployment
| System | A multi-GPU system reserved for you, sized to the model and your traffic |
|---|---|
| Commercial model | Hardware and managed service Quoted per deployment |
| Endpoint | OpenAI-compatible /v1 on your own machine, only your traffic |
| Location | Hosted in the EU: Málaga, Spain (eu-es-1) |
| Data handling | Zero prompt retention, same policy as the serverless API |
The model is validated on AxForge hardware and deployed on your system. After deployment, your endpoint speaks the OpenAI API with the model name you receive with the deployment.
Data & privacy
Prompts and completions are processed in memory in the EU, not written to disk, not logged, not retained, and never used to train anything. We keep only request metadata (token counts, timestamps, status) for billing and operations. The full policy is at axforge.ai/privacy.
FAQ
Not on the serverless API. It is available as a managed deployment: validated on AxForge hardware and deployed on a multi-GPU system for your traffic only. The serverless API serves Qwen3.8 27B.
A multi-GPU system of B200 class for the NVFP4 build, sourced against your request and confirmed in writing before anything is billed.
Super is the 120B mid-tier for balanced agent workloads on a single large-memory machine; Ultra is the flagship for the hardest tasks.
We publish only numbers we measure ourselves, and we haven't benchmarked this model on our nodes yet. For quality benchmarks, see the official model card.
A multi-GPU system reserved for you, running Nemotron 3 Ultra behind an OpenAI-compatible endpoint, hosted in the EU with zero prompt retention.
Hardware and the managed service are Quoted per deployment. Request deployment and we confirm both in writing.
Explore