# DeepInfra: what AI models say (September 2026)

Source: Memetik Index, https://www.memetik.ai/vendors/deepinfra
Website: https://deepinfra.com

DeepInfra was named in 8 of 50 AI answers across 1 category. Best position: Inference hosting, 16% answer share, ranked #13, named first in 0%.

## What DeepInfra is

DeepInfra provides AI inference infrastructure through APIs. It is for companies that need an inference provider within their AI stack, with access to more than 100 models and GPU instances.

It runs its own inference infrastructure in US-based data centres. Companies use its APIs to run AI inference, with pricing varying by model. Some language models are charged per token, while most other models are charged by inference execution time on a pay-as-you-go basis.

Pricing: Pricing varies by model: some language models are priced per token and most other models by inference execution time, with pay-as-you-go billing.
Founded: 2022
(From DeepInfra's own website, read 2026-09-02.)

## By category

| Category | Rank | Answer share | Named first | Models |
|---|---|---|---|---|
| Inference hosting (https://www.memetik.ai/index/inference-hosting) | #13 of 18 | 16% | 0% | GPT-5.6 Sol, Claude, Claude Fable 5, Gemini, Gemini 3.5 Flash |

## What the models said

> "The main players here are Together AI, Fireworks AI, Groq, Replicate, Deepinfra, and Hugging Face Inference."
> — Claude Opus 5, Inference hosting
