Model / inference API

DeepInfra

DeepInfra hosts open and some closed models behind an API, billed per token.

Website (opens in a new tab)

Checked 2026-11-05

DeepInfra hosts open and some closed models behind an API, billed per token. The pricing page we read on 5 November 2026 lists DeepSeek-V4-Flash at $0.09 per million input tokens and $0.18 output, Llama-3.3-70B-Instruct-Turbo at $0.10 and $0.32, and Kimi-K2.6 at $0.75 and $3.50. Third-party models appear too, such as claude-sonnet-4-6 at $3.00 and $15.00. A Priority tier costs 1.5x the base price and a Flex tier 0.8x. Custom deployments rent GPUs by the hour: A100 80GB at $1.49, H100 80GB at $2.90, H200 at $3.49 and B200 at $4.89, billed by the minute and invoiced weekly. We would point it at teams that want many open models on one pay-as-you-go bill. The limit we noticed is 200 concurrent requests per account unless you contact the company. The page mentions no free tier and says you add a card or prepay first. DeepInfra says it does not store submitted data on disk for standard inference and does not train on it, except for Google and Anthropic models, which follow those companies' policies. Bulk jobs may sit encrypted on disk for a short time. Facts come from the pricing page and data privacy docs we checked on 5 November 2026.

Mentioned here

  • Together AI

    Together AI serves open models by the token and also rents GPUs.

    • Model / inference API
    • OpenAI compatible

  • Fireworks AI

    Fireworks AI serves open models by the token and states zero data retention by default.

    • Model / inference API
    • OpenAI compatible

  • Groq

    GroqCloud serves open models on its own inference hardware, with speed as the draw.

    • Model / inference API
    • OpenAI compatible

  • OpenRouter

    OpenRouter routes more than 600 models through one API key.

    • Model gateway (router)
    • OpenAI compatible

  • Baseten

    Baseten is a hosted platform for model inference and training.

    • Model / inference API
    • OpenAI compatible

  • Hugging Face Inference

    Hugging Face Inference Endpoints deploys a model from the Hub onto dedicated, autoscaling hardware.

    • Model / inference API
    • OpenAI compatible

Questions

How much does DeepInfra cost?

Pay per token for hosted models, with image models priced per image. Each request can use a service tier: Standard is the base price, Priority is 1.5x for faster first tokens, and Flex is 0.8x for slower, non-production work. You can also deploy your own model on dedicated GPUs and pay for uptime (checked 5 November 2026).

Is DeepInfra OpenAI compatible?

Yes, it offers an OpenAI-compatible API, so most agents and SDKs can point at it by changing the base URL.

Reviews

Loading reviews…

Write a review

Your rating

Up to 1000 characters. Stars alone are enough.

Similar

  • Moonshot AI

    Moonshot AI sells the Kimi API through its developer platform, now at platform.kimi.ai, and it suits teams who want Kimi K3 and K3.1 Pro with a 1,048,576-token context window straight from the maker.

    • Model / inference API
    • OpenAI compatible

  • Featherless

    Featherless is a flat-rate subscription for open models, not a per-token meter, and it suits hobbyists and small teams who want to switch between many open models for one monthly price.

    • Model / inference API
    • OpenAI compatible

  • Z.AI

    Z.ai sells the GLM model family through its open platform API, and it suits developers who want GLM 5.3 straight from the maker, with free small models for testing.

    • Model / inference API
    • OpenAI compatible

  • Minimax

    MiniMax sells its model API through the MiniMax Open Platform, and it suits developers who want M3 or M3.1 Flash directly from the maker at a low token rate.

    • Model / inference API
    • OpenAI compatible

  • AI21

    AI21 sells the Jamba model family through an API, and it suits teams that want a long-context model from a vendor that says it does not train on customer inputs.

    • Model / inference API

  • Databricks

    Databricks sells model serving inside its data platform, so it suits teams that already keep data and pipelines on Databricks and want hosted open models next to them.

    • Model / inference API
    • OpenAI compatible