DeepInfra · Llama-4

Cloud PlatformOpenAI compatible

Serverless inference cloud for open-weight LLMs, image, video, TTS, transcription, embedding, and reranker models with an OpenAI-compatible API.

Active models
304
Families
79
Headquarters🇺🇸 Palo Alto, CA
Founded2022
API Base URL
Modelsdeepinfra.com/models
Capabilities
Model
Llama-4-Maverick-17B-128E-Instruct-FP8OSSmeta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8
Llama-4-Scout-17B-16E-InstructOSSmeta-llama/Llama-4-Scout-17B-16E-Instruct