AI API Pricing Explained
AI API pricing is easier to compare when every listing is reduced to the same basic signals: model, provider, input price, output price, source, and last checked date.
Canonical general AI API pricing page.
TLDR
Most LLM APIs price usage by input tokens and output tokens.
The cheapest listed price is not always the best fit; source, model quality, and rate limits still matter.
Use source links to confirm the latest provider pricing before a buying decision.
Who this is for
Product and engineering teams estimating model API costs.
Finance or procurement teams comparing official and marketplace listings.
Founders choosing a first model provider for production usage.
What AI API pricing means
An AI API lets a product send prompts or data to a model and receive a response. The bill usually depends on how much the model reads and how much it generates.
For text models, pricing is commonly shown per 1M input tokens and per 1M output tokens. That makes prices easier to compare across providers.
A token is a usage unit for model text. It is not a product bundle or a subscription by itself.
The fields worth comparing
A serious comparison should show more than a single price. Provider, model name, input price, output price, source link, verification status, and last checked date all change how useful a row is.
Latency and uptime are useful only when the provider publishes them. If those values are unknown, Inferras leaves them as N/A instead of filling fake numbers.
| Field | Why it matters |
|---|---|
| Input price | Important for search, classification, and long document analysis. |
| Output price | Important for chat, writing, coding, and agent workflows. |
| Source link | Lets you check the provider page behind the listing. |
| Last checked | Shows how fresh the public-source review is. |
A simple example
If one workflow sends long documents and asks for short summaries, input price may drive most of the cost. If another workflow produces long answers, output price may matter more.
That is why Inferras keeps input and output prices separate instead of showing one blended number.
Practical examples
Document review: high input usage, lower output usage.
Support assistant: moderate input usage, higher output usage.
Code generation: output price can become a large part of spend.
FAQ
AI API pricing
What should this general AI API pricing guide help me decide?
It helps you understand the core comparison fields before moving into model-specific pages, cheapest API pages, or provider due diligence.
When should I use the Price Radar instead of this guide?
Use the Price Radar when you already know the model or provider you want to compare and need current public listings with source links.
Where should I read more about token definitions?
Use the canonical AI token pricing guide for definitions and the input vs output token guide for billing mechanics.
Does this page rank providers?
No. It explains how to compare public AI API pricing fields without inventing rankings, ratings, or performance claims.
Source references
Related guides
2 likes
Comments
No approved comments yet
Reviewed comments will appear here.