Inferse
For vendors
AdvertiseTop spots, featured rows, home sponsor Get listedAdd a product we don't cover yet Claim a listingFree: verify the domain, then correct your page Vendor dashboardNumbers, orders and placements
List your platform
Vendor dashboard

Nemotron 3 Ultra (free)

NVIDIA· released 4 Jun 2026

nvidia/nemotron-3-ultra-550b-a55b:free
Input / 1M tokens
free
Output / 1M tokens
free
Context window1M tokens

Longest answer: 66K tokens

reads textopen weightstool callingreasoning
See it on openrouter.ai

About

Nemotron 3 Ultra (free) is NVIDIA’s open-weight reasoning and orchestration model, with 55 billion active parameters from 550 billion total in a mixture-of-experts design. It uses a hybrid architecture combining Transformer and Mamba components. This listing supports text input and output, reasoning, and tools. Its context window is 1,000,000 tokens, and it allows a maximum output of 65,536 tokens. The model weights are listed as open, with Hugging Face identifier nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16. The entry is marked free and carries the model identifier nvidia/nemotron-3-ultra-550b-a55b:free. It was released on June 4, 2026, and the official listing is on OpenRouter. For builders comparing configurations, the extended context and output limits distinguish the capacity details of this free listing; its stated inputs and outputs remain text, and the model includes tool support.

Who it is for

This listing may suit teams seeking free access to an open-weight, text-focused reasoning model with tool support. Its 1,000,000-token context and 65,536-token output limit may matter for longer workflows.

What is good

  • Marked free and open weights
  • 1,000,000-token context window
  • Maximum output of 65,536 tokens
  • Reasoning and tool support

What to know first

  • Listed inputs and outputs are text only
  • 55B active within 550B total parameters

Verdict

The free listing pairs open weights with reasoning and tool support. It specifies a 1,000,000-token context window and a maximum output of 65,536 tokens, with text-only inputs and outputs.

Details

Lab
NVIDIAopenrouter.ai · 3 Oct 2026
Context
1,000Kopenrouter.ai · 3 Oct 2026
Input price
Freeopenrouter.ai · 3 Oct 2026
Output price
Freeopenrouter.ai · 3 Oct 2026
Max output
65,536 tokensopenrouter.ai · 3 Oct 2026
Inputs
textopenrouter.ai · 3 Oct 2026
Open weights
Yesopenrouter.ai · 3 Oct 2026

More newest models

See the list
Ling 3.1 Flash ctx262KinFreeoutFree inclusionAIApodex 1.1 Mini (free) ctx262KinFreeoutFree ApodexPareto 26.10 Preview ctx1,049Kin$0.8 / 1Mout$3.2 / 1M unbiasedGPT-6.1 Sol ctx1,050Kin$2 / 1Mout$10 / 1M OpenAIGPT-6.1 Sol Pro ctx1,050Kin$2 / 1Mout$10 / 1M OpenAIClaude Sonnet 5.5 ctx1,000Kin$2 / 1Mout$10 / 1M Anthropic

Listed on Inferse

Sources