NVIDIA Nemotron 3 Ultra NVFP4

NVIDIA frontier-scale hybrid LatentMoE (550B params, 55B active) interleaving Mamba-2 and MoE layers, for multi-step agents and long-context reasoning.

262K context·$0.60 in·$2.40 out·Jun 2026 release·Reasoning, Tools / functions

Available on

Fireworks AI

Providers

Where NVIDIA Nemotron 3 Ultra NVFP4 runs in Big-AGI, with each service's own model ID and USD-per-1M-token rates. The summary above prefers the creator's claim; rates below are each service's own.

ServiceModel IDContextInputOutputCache read

Fireworks AI

accounts/fireworks/models/nemotron-3-ultra-nvfp4

262K

$0.60

$2.40

$0.12

Specs as reported by each service · usage over the last 2 days · refreshed hourly