DeepInfra Pricing
deepinfra.com · AI Inference · refreshed 2026-05-03deepinfra.com/pricing
Simple Pricing, Deep Infrastructure
DeepSeek V3.1 $0.21/$0.79 per MTok; Llama 3.1 8B $0.02/$0.05; H100 custom-LLM $1.79/GPU-hr
$0
Self-serve
Self-serve
Self-serve
Self-serve
Contact sales
Plans & what each tier includes
Synthesized from DeepInfra's public pricing page on 2026-05-03. Tier names normalized to a starter / team / business / enterprise pattern; the actual DeepInfra pricing page is the source of truth for in-tier features.Free
$0
foreverTier 1 starts with $20 invoicing threshold, PAYG
Access to DeepInfra's ai inference surface
Standard per-unit pricing applies once free quota is consumed
Full public API documentation available
Scale
Self-serve
higher tierHigher monthly quotas and per-unit discounts
Reduced per-unit cost vs. Entry tier
Priority support, additional seats, advanced features
Card checkout, no sales contact required
Enterprise
Contact sales
customCustom pricing negotiated with the DeepInfra sales team
Dedicated account manager and SLA commitments
Volume discounts on per-unit costs
Procurement-friendly: invoicing, NDAs, security questionnaires
What DeepInfra charges
DeepInfra operates in the ai inference space. The published pricing page lists 5 tiers, including a free tier, with per-unit pricing of DeepSeek V3.1 $0.21/$0.79 per MTok; Llama 3.1 8B $0.02/$0.05; H100 custom-LLM $1.79/GPU-hr. The pricing page positions plans as: "Simple Pricing, Deep Infrastructure".
For the deeper analysis — score, simulation, pros & cons for AI agents, and the one-line fix that would lift this listing into the Verified band — see the agent-readiness scorecard.DeepInfra pricing FAQ
DeepInfra offers a free tier (Tier 1 starts with $20 invoicing threshold, PAYG). Paid pricing is sales-led — the live pricing page does not list a self-serve paid plan as of 2026-05-03.
Yes — DeepInfra publishes a free tier: Tier 1 starts with $20 invoicing threshold, PAYG. This makes it possible to evaluate DeepInfra without committing budget.
DeepInfra's public pricing page lists the following per-unit pricing: DeepSeek V3.1 $0.21/$0.79 per MTok; Llama 3.1 8B $0.02/$0.05; H100 custom-LLM $1.79/GPU-hr. This is the rate captured by xpay's Agent-Ready SaaS Index on 2026-05-03 and is the basis for the agent-readiness score on the linked scorecard.
DeepInfra's pricing page positions itself as: "Simple Pricing, Deep Infrastructure". Per-unit pricing of DeepSeek V3.1 $0.21/$0.79 per MTok; Llama 3.1 8B $0.02/$0.05; H100 custom-LLM $1.79/GPU-hr is published. For the strict definition of agent-buyable PPU — a per-task purchase that an AI agent can complete without an annual commitment — see the agent-readiness scorecard.
Partially. 5 DeepInfra tiers are self-serve via card checkout; 1 tier requires contacting sales.
Yes — DeepInfra's API documentation is publicly accessible at https://docs.deepinfra.com. An AI agent or developer can inspect the integration surface without authentication.
xpay scored DeepInfra 76 / 100 on its Agent-Ready SaaS Index, which places it in the Approaching Agent-Ready band. Based on public pricing signals, xpay verified that DeepInfra exposes per-MTok rates for dozens of model families, per-image formulas for FLUX, and per-GPU-hour custom LLM rates on a public anonymous page. Self-serve usage tiers auto-promote on spend. xpay could not locate /.well-known/ai-pricing.json or anonymous agent purchase paths. For the full 6-step purchase simulation and the one-line fix that would lift this listing into the next band, see the linked scorecard.
In the AI Inference category, xpay's Agent-Ready Index lists multiple alternatives ranked by their ability to support AI agent purchase. Visit the agent-readiness scorecard for DeepInfra to compare with 3 nearest peers, or browse the full AI Inference category from the index.
Compare with similar SaaS
Can an AI agent buy from DeepInfra?
76
/ 100 in xpay's Agent-Ready SaaS IndexBased on public pricing signals, xpay verified that DeepInfra exposes per-MTok rates for dozens of model families, per-image formulas for FLUX, and per-GPU-hour custom LLM rates on a public anonymous …
See the full scorecardIncludes the 6-step agent purchase simulation, score breakdown across 7 dimensions, and the one-line fix that would lift DeepInfra into the Agent-Ready Verified band.At a glance
Tier 1 starts with $20 invoicing threshold, PAYG
Free
DeepSeek V3.1 $0.21/$0.79 per MTok; Llama 3.1 8B $0.02/$0.05; H100 custom-LLM $1.79/GPU-hr
5 tiers
5
Public docs
Compare nearby
DeepInfra pricing summary derived from a live read of https://deepinfra.com/pricing on 2026-05-03. For the agent-readiness assessment — including the 6-step purchase simulation and a per-dimension scorecard — see xpay's Agent-Ready SaaS Index entry for DeepInfra.