Nemotron 3 Ultra

Nemotron 3 Ultra

Developer	NVIDIA
Release date	June 4, 2026
Family	Nemotron
Generation	Nemotron 3
Size class
Context window	262K tokens
Max output	183K tokens
Input	Text
Output	Text
Knowledge cutoff	2025-09
Weights	Open weights
Website	nvidia.com
Links	Model card Hugging Face

input: $0.60
cached: $0.12
output: $2.40
Context: 262Ktokens
Released: 5mo ago
Size: 550Bparameters

Specifications

Model

Developer: NVIDIA
Family: Nemotron Nemotron 3
Type: Language model
Released: June 4, 2026
Model ID: nvidia/nemotron-3-ultra-550b-a55b

Capabilities

Input → output: Text → Text
Native context: 262K tokens
Max output: 183K tokens
Tokenizer: Other
Knowledge cutoff: 2025-09

Weights

Openness: Open weights
Size: 550B
Hugging Face: nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16

Sources:models.devOpenRouter

Metadata synced Nov 4, 2026, 7:02 PM UTC

To give you the best experience, we use cookies for personalization, insights, and performance. Learn more.