o4-mini
o4-mini by OpenAI. Official list price: $1.1 per 1M input and $4.4 per 1M output tokens.
Official price benchmark
USD per 1M tokens
- input
- $1.10
- output
- $4.40
- cache read
- $0.275
Your budget
Input:output ratio · Similar amounts of input and output
Purchasing power ranking
$100 · o4-mini · Balanced
Best price
RunAPI
−91%
providers
1
max −91%
Purchasing power ranking
The provider's published prices next to the vendor's official list.
Some links on this site are partner links. They do not affect prices or ranking. Uptime and latency come from our own GET /models probes every 30 minutes, not a guarantee that the model is available.
Cost estimate
What one request and a month of traffic cost at the selected provider.
per request
$0
per month
$1.87
RunAPI · −91% · 1.20K / 500
Other models
Frequently asked questions
Answers are generated from the live quotes.
We collect o4-mini prices two ways: the official OpenAI list and quotes from providers our crawler checks daily. Providers in our catalogue sell o4-mini at $0.096 in / $0.384 out (91% below the official list). Everything is normalised to a blended cost at a chosen input:output ratio, so “o4-mini price per 1M tokens” means the same thing on every page.
What does the o4-mini API cost?
The official OpenAI list price is $1.10 per 1M input tokens and $4.40 per 1M output tokens. Cached input is $0.275 per 1M tokens, 4× cheaper than the regular input.
How much does 1 million tokens of o4-mini cost?
1M input tokens cost $1.10, 1M output tokens — $4.40. With a 1:1 input/output load a million tokens works out to about $2.75. Chatbots with long context push the blended price down; code generation pushes it up.
Where is the o4-mini API cheapest?
Our measurements show RunAPI as the cheapest: 91% below the official price. Quotes from 1 providers refresh daily — see the table above.
What is the o4-mini context window?
The context window is 200K tokens. Some vendors raise the price above a context threshold — check the plan notes.
How do I check that a provider key works?
The free checker sends a GET request to your endpoint's model list and reports status and latency. The key is not stored server-side.