Ficelle
Nemotron 3 Ultra free API: where to use it for free
NVIDIA Nemotron 3 Ultra 550B A55B is free on OpenRouter and Requesty. The exact model ids, free-tier limits, and how to fail over when one runs out.
Checked against NVIDIA's and each provider's official pages
Where Nemotron 3 Ultra is free
NVIDIA publishes the weights under OpenMDW 1.1, and two providers serve it on a free tier.
| Provider | Model id | Free-tier terms |
|---|---|---|
| OpenRouter | nvidia/nemotron-3-ultra-550b-a55b:free |
20 requests per minute, 50 or 1,000 per day. 1,000,000-token context. |
| Requesty | nvidia/nemotron-3-ultra-550b-a55b |
200 requests a day and 20 a minute for new organizations, shared across all free models. |
What Nemotron 3 Ultra is
Nemotron 3 Ultra is NVIDIA's largest open Nemotron 3 model, a 550B mixture-of-experts with 55B active parameters, built for complex multi-step agents, long-context analysis and high-accuracy reasoning over code, math and science.
- Parameters
- 550B total, 55B active (MoE)
- Context window
- Up to 1M tokens
- Input
- Text
- Output
- Text
- Reasoning
- Switchable on or off
- Languages
- 10, including English, French, German, Hindi, Japanese and Chinese
Before you rely on it
- A large model on a free tier can be slower or more often at capacity than smaller ones; failover covers the gaps.
- Requesty says its free models are free for now and that pricing changes will be announced in its changelog.
Use Nemotron 3 Ultra with automatic failover
A single free tier runs out. Ficelle puts the free models of every provider you
configure behind one local OpenAI-compatible endpoint. Call a routing profile and
Nemotron 3 Ultra is one of the models it can pick; when it rate-limits or fails, the next
eligible free model answers. To pin Nemotron 3 Ultra alone, call the id that
ficelle models lists for it, knowing a pinned model has no failover.
from openai import OpenAI
client = OpenAI(base_url="http://127.0.0.1:8646/v1", api_key="ficelle-local")
# A routing profile keeps failover: Nemotron 3 Ultra is one of the free
# models it can pick, and the next eligible one answers if it fails.
client.chat.completions.create(
model="ficelle/auto-reasoning",
messages=[{"role": "user", "content": "..."}],
)Routing profiles that fit what Nemotron 3 Ultra is built for:
- Reasoning
ficelle/auto-reasoning - Agent orchestration
ficelle/auto-orchestrator - Long context
ficelle/auto-long
Whether a profile routes to Nemotron 3 Ultra on your install depends on your providers' live catalogs and Ficelle's eligibility checks, decided at request time.
Questions
Is Nemotron 3 Ultra free to use?
NVIDIA publishes the weights under OpenMDW 1.1, and two providers serve it on a free tier. When we checked on 28 September 2026, OpenRouter and Requesty served it at no charge, each under its own free-tier limits.
Where can I use the Nemotron 3 Ultra API for free?
OpenRouter as nvidia/nemotron-3-ultra-550b-a55b:free; Requesty as nvidia/nemotron-3-ultra-550b-a55b. Free catalogs change often, so check the provider's own page before you rely on one.
What is Nemotron 3 Ultra?
Nemotron 3 Ultra is NVIDIA's largest open Nemotron 3 model, a 550B mixture-of-experts with 55B active parameters, built for complex multi-step agents, long-context analysis and high-accuracy reasoning over code, math and science.
Can I call Nemotron 3 Ultra through Ficelle?
Yes, when one of your configured providers serves it for free. Run ficelle models to see the ids your install can route to. Calling that id pins the one model, with no failover; calling a routing profile such as ficelle/auto-reasoning keeps failover, so another free model answers when Nemotron 3 Ultra is rate-limited.
Sources
Official pages this guide rests on, read on 28 September 2026:
