DeepInfra

Low-cost hosted inference for open models.

Visit site
Proprietary usage
LLM provider
In these stacks
Write code with AIBuild my own agent
LicenseProprietary
PricePer-token
UsageMetered
Model supportmulti
Languages
GitHub stars
Last activity
Verified2026-06-27
Verified byseed

A cost-focused inference platform hosting open-weight models via an OpenAI-compatible API at low per-token prices.

Strengths

  • Among the cheapest per-token pricing for open models
  • OpenAI-compatible drop-in API

Tradeoffs

  • Proprietary hosted platform
  • Less enterprise polish than larger providers