DeepInfra

Overview

DeepInfra provides a platform for running AI models in the cloud, focusing on ease of use, scalability, and cost-effectiveness. It offers a simple API (REST, Python, JavaScript) and supports OpenAI API compatibility for easy migration. The platform handles servers, GPUs, and scaling, allowing developers to concentrate on their applications. DeepInfra features pay-as-you-go pricing, charging only for input and output tokens for LLMs or inference execution time for other models. The platform supports over 100 models, offering options for text generation, image creation, video processing, and speech recognition. It also provides zero retention policy, ensuring data privacy and compliance with SOC 2 and ISO 27001 standards. DeepInfra leverages its own inference-optimized infrastructure in US-based data centers for performance and reliability.

Common tasks

Text Generation Text-to-Image Text-to-Video Automatic Speech Recognition Embeddings

FAQ

View all

What is DeepInfra?

DeepInfra is a cloud platform that simplifies running machine learning models by providing developer-friendly APIs, handling servers, GPUs, and scaling.

How does DeepInfra pricing work?

DeepInfra uses a pay-as-you-go model. For LLMs, you pay per input and output token. For other models, you are billed for inference execution time.

What kind of models are available on DeepInfra?

DeepInfra provides 100+ models including text generation, text-to-image, text-to-video, and speech recognition models. Models from DeepSeek, Qwen, Meta, and NVIDIA are available.

Is DeepInfra secure?

Yes, DeepInfra is SOC 2 and ISO 27001 certified, uses a zero-retention policy, and follows best practices in information security and privacy.

FAQ+