Nemotron-H 56B

Training compute

6.7×10²⁴ FLOP

Parameters

56B

Published

Apr 14, 2025

Nemotron-H 56B is an AI model developed by NVIDIA (United States), first published in April 2025. It works in the language domain, on tasks such as language modeling/generation, question answering, translation and 2 more.

Training it took an estimated 6.7×10²⁴ FLOP of compute (estimation method: operation counting). The model has 56,000,000,000 parameters. Training ran on 6,144 NVIDIA H100 SXM5 80GB.

Access: Open weights (non-commercial). Its weights are openly available. Epoch AI rates the confidence of this record as confident.

Full record

Organization: NVIDIA
Country of organization: United States
Domain: Language
Task: Language modeling/generation, Question answering, Translation, Quantitative reasoning, Code generation
Training compute: 6.7×10²⁴ FLOP
Compute estimation method: Operation counting
Parameters: 56,000,000,000
Training hardware: NVIDIA H100 SXM5 80GB
Chips used: 6,144
Training power draw: 8.4 MW
Numerical format: FP8
Model accessibility: Open weights (non-commercial)
Open weights: Yes
Epoch confidence: Confident