Llama-3.1-Nemotron-Ultra-253B-v1

Open weights Active Text LLM Dense llama-3-1-nemotron-ultra-253b-v1

253B reasoning model derived from Llama 3.1 405B via neural architecture search; fits one 8xH100 node.

Released
Apr 7, 2025
Context window
131K tokens
Max output
Not disclosed
Parameters
253B
Input price
Not disclosed
Output price
Not disclosed

Prices are first-party list prices in USD at the standard tier. Last verified Oct 30, 2026.

Pricing

No first-party API. This is an open-weights release you can run yourself.

Specifications

Developer
NVIDIA
Model type
Text LLM
Architecture
Dense
Parameters
253B
Input
Text
Output
Text
Knowledge cutoff
Not disclosed
License
NVIDIA Open Model License
API access
No first-party API
Capabilities
Flagship Long context Reasoning Tool use