Compare AI models

Specifications
Parameters
—550B
Context window
128k1M
API pricing
Cheapest input
—$0.40DeepInfra
Cheapest output
—$1.76DeepInfra
—$210KT
Benchmarks
GPQA Diamond
76.26%86.7%
Overview
CompanyMistralNVIDIA
Release dateSep 17 2025Jun 4 2026
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Nemotron 3 Ultra leads Magistral Medium 1.2 on 1 of the 1 benchmark they both report (GPQA Diamond). Magistral Medium 1.2 shipped 260 days before Nemotron 3 Ultra, so benchmark comparisons should account for the intervening progress.

Context windows are 128k (Magistral Medium 1.2) vs 1M (Nemotron 3 Ultra). Magistral Medium 1.2 is closed, while Nemotron 3 Ultra is open weight.

On GPQA Diamond, Nemotron 3 Ultra leads at 86.7% vs Magistral Medium 1.2 at 76.26%.