⚡ Ultra Low-Cost Serverless Inference

DeepInfra API Pricing & Rates

The developer favorite for bare-metal cost efficiency. DeepSeek V3 at $0.14/$0.28, Whisper audio at $0.0005/min, and zero cold starts.

DeepSeek V3 (In/Out)
$0.14 / $0.28
Per 1M tokens serverless
DeepSeek R1 (Reasoning)
$0.55 / $2.19
Full 671B reasoning model
Whisper Large v3
$0.0005
Per minute (91% off OpenAI)
FLUX.1 Schnell
$0.003
Per generated image

🧮 DeepInfra Monthly Bill Calculator

Calculate your blended multi-modal API spend on DeepInfra:

DeepSeek V3 Tokens (Millions / month) 50M tokens
Audio Transcription (Hours / month) 200 hours
LLM Inference Bill
$10.50
Whisper Audio Bill
$6.00
DeepInfra Total
$16.50
OpenAI Equivalent
$197.00

📋 DeepInfra Verified Model Rate Card

Model Input / 1M Output / 1M Context Throughput
deepseek-ai/DeepSeek-V3 $0.14 $0.28 64k ~60 tok/s
deepseek-ai/DeepSeek-R1 $0.55 $2.19 64k ~45 tok/s
meta-llama/Llama-3.3-70B-Instruct $0.23 $0.40 128k ~72 tok/s
Qwen/Qwen2.5-Coder-32B-Instruct $0.18 $0.35 32k ~85 tok/s
openai/whisper-large-v3 $0.0005 / min - - Real-time 15x

💻 Python OpenAI SDK Drop-in

from openai import OpenAI client = OpenAI( base_url="https://api.deepinfra.com/v1/openai", api_key="your_deepinfra_api_key" ) response = client.chat.completions.create( model="deepseek-ai/DeepSeek-V3", messages=[{"role": "user", "content": "How do I optimize SQL queries for BigQuery?"}] ) print(response.choices[0].message.content)

Frequently Asked Questions

What are DeepInfra's token prices for popular open-weights models?
DeepInfra prices DeepSeek V3 at $0.14/1M in and $0.28/1M out; DeepSeek R1 at $0.55/1M in and $2.19/1M out; Llama 3.3 70B Turbo at $0.23/1M in and $0.40/1M out; and Qwen 2.5 Coder 32B at $0.18/1M in and $0.35/1M out.
How does DeepInfra Whisper pricing compare to OpenAI?
DeepInfra charges $0.0005 per audio minute ($0.03 per hour) for Whisper Large v3, which is 91.7% cheaper than OpenAI's official Whisper API ($0.006 per minute or $0.36 per hour).
Can I use the OpenAI Python SDK with DeepInfra?
Yes, setting base_url='https://api.deepinfra.com/v1/openai' allows full drop-in compatibility with standard OpenAI API clients.