All LLM provider verticals

GPT-4o, GPT-4 Turbo, o1 reasoning models

Claude Sonnet 4.5, Opus 4.6, Haiku 4.5

Gemini 2.5 Pro, 2.5 Flash, Flash-8B

Mistral Large, Codestral, Pixtral

Llama 3.3, Llama 3.2 Vision, Llama 4

Grok-3, Grok-2 Vision

Sub-100ms inference for Llama, Mixtral, Whisper

200+ open-weights models — Llama, Qwen, DeepSeek

Fast inference for Llama, DeepSeek, function calling

Sonar Pro — web-grounded answers with citations

Command R+, Embed v3, Rerank

Jamba 1.5 Large — hybrid Mamba-Transformer

Claude, Llama, Titan, Mistral — through your AWS account

GPT-4o on Azure — your Azure subscription, your data residency

Gemini + Claude + Llama on GCP — your project, your residency

300+ models across providers — universal router

Cheap open-weights inference — Llama, Qwen, Mixtral

Run any open-weights model — Llama, FLUX, Whisper, ComfyUI

Wafer-scale inference — fastest tokens/sec on the market

GPU-cloud inference — Hermes 3, Llama 3.3

Optimized inference — Llama, Stable Diffusion, custom

Production-grade open-weights inference

NVIDIA-optimized microservices for open-weights LLMs

DBRX, Llama, Mixtral served on Databricks

Hybrid Mamba-Transformer with 256K context — open weights