DeepInfra
Hosted inference API for open models.
Key features
- APIs for text, speech, image, music, and video models
- Hourly on-demand rental of NVIDIA DGX B300 GPUs
- Language models billed per token; others by runtime
- Zero retention of inputs, outputs, and user data
- SOC 2 and ISO 27001 certified US data centers
- Category
- LLM APIs
- Who it's for
- Startups and enterprises that need hosted model inference.
- Tags
- apiinferencellm-apiopen-modelsgpuhostingvoice-speechimage-videosecurity
- Website
- deepinfra.com
- Listed
- 2026-09-15
Sourced notes
Frequently asked questions
Is DeepInfra free?
No. DeepInfra is a paid product with no free plan listed on its site.
What is DeepInfra used for?
Hosted inference API for open models. Typical uses: aPIs for text, speech, image, music, and video models; hourly on-demand rental of NVIDIA DGX B300 GPUs; language models billed per token; others by runtime.
What are the best DeepInfra alternatives?
Alternatives in the same category include fal, Fireworks AI, Mistral, Replicate, Together AI.
Related in LLM APIs
See all →- Fireworks AI Inference platform for open models.
- Groq Inference API known for low-latency open models.
- Together AI Inference and fine-tuning API for open models.
- Amazon Bedrock AWS API for foundation models, including Anthropic and Amazon models.
- Anthropic Console Console and API for Claude models.
- Azure OpenAI OpenAI models hosted on Microsoft Azure.
