ModelPricing.aio3 API Pricing
o3 costs $2.00 per million input tokens and $8.00 per million output tokens on the OpenAI API. Prices are tracked from the official OpenAI pricing table and used by the ModelPricing.ai estimation API, so this page and the API always quote the same numbers. See all OpenAI model pricing or the full LLM pricing comparison.
o3 Pricing Per Million Tokens
| Tier | Input / 1M | Output / 1M | Cache read / 1M | Cache write / 1M |
|---|---|---|---|---|
| Standard | $2.00 | $8.00 | $0.500 | — |
What o3 Costs In Practice
Monthly cost at three representative workloads, using the standard text rate in effect today.
| Workload | Monthly cost |
|---|---|
| Light — 1M in / 200K out | $3.60 |
| Moderate — 10M in / 2M out | $36.00 |
| Heavy — 100M in / 20M out | $360.00 |
Need your own numbers? Use the LLM cost calculator or call the estimation API directly.
Compare o3 With Other Models
Looking for something cheaper from OpenAI? GPT-5 Nano is the lowest-priced option on input tokens.
o3 Pricing FAQ
How much does o3 cost per token?
o3 costs $2.00 per million input tokens and $8.00 per million output tokens on the OpenAI API. That works out to $0.0020 per thousand input tokens.
How much does it cost to run o3 per month?
At 1M in / 200K out per month, o3 costs about $3.60. A heavier workload of 100M input and 20M output tokens costs about $360.00.
Does o3 support prompt caching?
Yes. Cached input tokens are billed at $0.500 per million on o3, compared with $2.00 per million for uncached input.
Estimate o3 costs programmatically
POST a token count to /v1/estimate with model o3 and get a priced breakdown back, promotions
already applied.