ProvidersNVIDIA

Model provider

NVIDIA

NIM-hosted Nemotron and open models optimized for accelerated inference.

chat

Try in Playground Provider website

Available through one endpoint

5 NVIDIA models

API documentation →

NVIDIANemotron 3 Ultranemotron-ultra

NVIDIA flagship 550B MoE reasoning model served through NIM.

128K context$0.80 / $3.20reasoningopen-weights NVIDIANemotron 3 Supernemotron-3-super

Mid-size NVIDIA Nemotron MoE for reasoning and agents on NIM.

128K context$0.25 / $1.00reasoningtoolsopen-weights NVIDIANemotron 3 Nanonemotron-3-nano

Low-cost NVIDIA Nemotron MoE for fast chat and extraction on NIM.

128K context$0.06 / $0.24fastlow-costopen-weights NVIDIALlama Nemotron Super 49Bnemotron-super-49b

Llama-based Nemotron tuned for reasoning, served through NIM.

128K context$0.12 / $0.36reasoningopen-weights NVIDIAGPT-OSS 120B on NVIDIAnvidia-gpt-oss-120b

OpenAI open-weights 120B served through NIM.

128K context$0.10 / $0.40chatcodeopen-weights