Model Terminal

NVIDIA Nemotron 3 Ultra (free)

Nemotron 3 Ultra is NVIDIA's largest open reasoning and agentic large language model, featuring 550B total parameters with 55B active via a hybrid Transformer/Mamba mixture-of-experts architecture. It is designed for long-running agent workflows such as planning, tool use, coding, and deep research, supporting up to 1M-token context windows. Source: supabase.

Value score
Not scored
Context
tokens
Max output
tokens
Price

Capability radar

Peer value bars

Identity

Developer
NVIDIA
Openness
Modalities
Release
Knowledge cutoff
Deprecation
API docs

Benchmarks

No benchmark scores yet.

Compare nearby

Pricing

Input / 1M
Output / 1M
Speed