GPUy.dev
DeepInfra logo

How to Get Started on DeepInfra: API Keys, Credits & Usage (2026)

Beginner

Estimated setup time: 2 minutes•Updated for 2026 infrastructure

Start on DeepInfra
KYC Policy
None (Zero KYC)
Payment Methods
Credit / Debit Card • Prepaid Balance
Min. Deposit
$0 (Free tier included)
Billing Model
Strictly pay-per-token / pay-per-second serverless
Trial Credits
$1.80 instant free credit on signup
Terminal Access
Serverless API Only (No SSH needed)

What You Need Before Starting

✓A GitHub, Google, or email account

Step-by-Step Setup Walkthrough

3 Steps to First Run
1

Instant 1-Click Signup with GitHub or Google

Step 1 of 3

Get instant access without waiting for approval.

  • •Go to deepinfra.com via GPUy's direct link.
  • •Log in with GitHub or Google. Your account is immediately created with $1.80 in free credits to test all open-source models.
Pro Tip: No credit card is required to use the initial $1.80 free balance.
2

Generate Your API Key

Step 2 of 3

Create your bearer token for API integration in 5 seconds.

  • •Click 'API Keys' in the top navigation.
  • •Click 'Create New Key', give it a label, and copy the secret token.
3

Make Your First Request with OpenAI SDK

Step 3 of 3

Point your existing OpenAI code to DeepInfra's high-speed inference endpoint.

  • •DeepInfra supports chat completions, embeddings, and image generation (FLUX.1 / SDXL).
Python Drop-In Example
python
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_DEEPINFRA_KEY",
    base_url="https://api.deepinfra.com/v1/openai"
)

chat = client.chat.completions.create(
    model="meta-llama/Llama-3.3-70B-Instruct",
    messages=[{"role": "user", "content": "Why is GPUy the best compute index?"}]
)
print(chat.choices[0].message.content)

Insider Cost-Saving & Anti-Waste Tactics

#1: Enable streaming on token completions to immediately terminate generation when satisfactory output is received.

Frequently Asked Questions (DeepInfra Onboarding)

Q:Does DeepInfra charge while idle?

No. DeepInfra is 100% serverless. You are only billed for the exact tokens generated or images created. There are zero idle instance or storage fees.

Q:How cheap is Llama 3.3 70B on DeepInfra compared to OpenAI?

DeepInfra charges ~$0.23 per million input tokens and ~$0.40 per million output tokens, which is over 80% cheaper than proprietary models like GPT-4o.

Official Developer Onboarding Channel

Ready to Launch on DeepInfra?

Followed the guide? Jump directly to the official console using our verified partner route with guaranteed zero markup and active referral tracking.

Compare Live RatesGo to DeepInfra Console