Trust
Security & compliance
SOC 2ISO 27001
SLA uptime target: Not specified
Security review turnaround: Not specified
DPA available: Not specified
Serverless inference platform for running popular open-source AI models with per-token pricing and low latency.
Pricing
From $0.02 / 1M tokens
Founded
2022
Team size
11-50 employees
Headquarters
San Francisco, CA
Product Preview
Preview from deepinfra.com
At a glance
Best for
Startup (1-50), Small business (51-200), Mid-market (201-1,000), Enterprise (1,000+)
Pricing model
Usage-based
Try before you buy
Integrates with
Claude (Anthropic)DeepSeekLlama (Meta)Qwen (Alibaba)Gemini (Google)+4
Security
SOC 2ISO 27001
About DeepInfra
DeepInfra is a serverless AI inference cloud that runs open-source and frontier LLMs, embedding, image, and audio models behind a simple OpenAI-compatible API, billed per token or per second of GPU time. It also rents dedicated GPUs (A100/H100/H200) for custom or fine-tuned models. Its users are developers and AI teams that want low-cost, on-demand model hosting without managing infrastructure.
Buyer Fit & Positioning
Procurement & Fit
Structured facts from the vendor to help your security, finance, and procurement reviews move faster.
Pricing
Commercial model
Pricing model: Usage-based
Free trial: Yes
Free plan: No
Contract minimum: Not specified
Procurement
Purchasing & legal
The vendor hasn’t added purchasing & legal details yet.
Fit
Best-fit company size
Startup (1-50)Small business (51-200)Mid-market (201-1,000)Enterprise (1,000+)
Implementation & Procurement
Commercial Fit & Ecosystem
Proof, Outcomes & Momentum
Alternatives, Migration & Buyer Objections
Buy through Cubbie
DeepInfra
DeepInfra doesn’t have a live offer on Cubbie yet, but you can still buy it through us. Request it and Cubbie’s procurement team will source pricing, negotiate terms, and set up the purchase for you.
Want a negotiated price? Request a quote for DeepInfra and Cubbie routes it to the vendor and our procurement team.
