DeepInfra
Low-cost, OpenAI-compatible API for popular open models.
A serverless inference provider hosting a broad catalog of open-weight models behind an OpenAI-compatible API, known for aggressive per-token pricing; also offers dedicated GPU deployments.
- Vendor
- DeepInfra
- Category
- Inference & model hosting
- Pricing
- Paid
- Open source
- No
- Platforms
- API
- Website
- deepinfra.com
Features
- OpenAI-compatible API
- Large open-model catalog
- Dedicated deployments
Best for
- High volume / low cost
Works with
- gpt-oss-120b — OpenAI. OpenAI’s Apache-2.0 open-weight reasoning model; runs on a single 80GB GPU.
- Llama 4 Scout — Meta. Efficient open Llama 4 model with a very long native context.
- Qwen3.6 35B A3B — Alibaba (Qwen). Open MoE with just 3B active parameters — popular for local inference.
- DeepSeek V4.1 Flash — DeepSeek. Open-weight MoE built on DeepSeek’s new encoder-decoder architecture.
More inference & model hosting
- Ollama — Ollama. Run open models locally with one command.
- vLLM — vLLM project. High-throughput LLM serving engine.
- llama.cpp — ggml. LLM inference in C/C++ on any hardware.
- Amazon Bedrock — AWS. Foundation models on AWS.
- Vertex AI — Google Cloud. Google Cloud’s AI platform.
- LM Studio — Element Labs. Desktop app to discover and run local LLMs.
- Azure AI Foundry — Microsoft. Build and run AI apps on Azure.
- Groq — Groq. Ultra-fast inference on custom LPUs.