OpenAI o3 API pricing at a glance
OpenAI lists o3 standard API pricing at $2.00 per 1 million input tokens, $0.50 per 1 million cached input tokens, and $8.00 per 1 million output tokens. These are API token rates, not ChatGPT subscription prices. (developers.openai.com)
Important for production planning: OpenAI marks o3 and the snapshot o3-2025-04-16 as deprecated and identifies GPT-5 as its successor. Confirm availability for your account and migration implications before building a new production dependency around o3. (developers.openai.com)
| o3 API processing | Input per 1M tokens | Cached input per 1M tokens | Output per 1M tokens |
|---|---|---|---|
| Standard | $2.00 | $0.50 | $8.00 |
How to estimate an o3 API bill
For standard processing, estimate token charges with:
cost = (uncached input tokens ÷ 1,000,000 × $2.00) + (cached input tokens ÷ 1,000,000 × $0.50) + (output tokens ÷ 1,000,000 × $8.00)
This example assumes a month of 800,000 uncached input tokens, 200,000 cached input tokens, and 250,000 total output tokens:
- Uncached input: 0.8 × $2.00 = $1.60
- Cached input: 0.2 × $0.50 = $0.10
- Output: 0.25 × $8.00 = $2.00
- Estimated total: $3.70 in token charges for those assumed volumes.
The Responses API reports total output usage as output_tokens. For reasoning models such as o3, reasoning tokens are included in output usage and billed at the output-token rate; don’t add a separate reasoning-token estimate to that total. The example’s 250,000 output tokens therefore includes any reasoning tokens. Use the reported usage, not visible answer length alone, when forecasting spend. (developers.openai.com)
This is a hypothetical calculation, not a measured result. It excludes any separately priced tools or other charges. Actual spend depends on billable token volumes in your workload.
When might o3’s cost make sense?
OpenAI describes o3 as a reasoning model for complex, multi-step tasks. It may be worth evaluating where a workload requires substantial reasoning, but whether its capabilities justify the cost depends on your results, latency needs, and token usage. (developers.openai.com)
For a new production system, o3’s deprecated status is a significant trade-off even if its listed rates fit your budget. OpenAI identifies GPT-5 as its successor. Compare currently supported models against your product’s requirements, and test representative workloads when quality or latency matters; a lower rate alone does not prove another model will meet your targets. (developers.openai.com)
Bottom line: OpenAI lists o3 standard API rates of $2 per million input tokens, $0.50 per million cached input tokens, and $8 per million output tokens. Forecast from reported total token usage—including reasoning tokens in output—and verify availability and migration implications before adopting deprecated o3 for new production work.