Model positioning
OpenAI o3 is a reasoning model for complex tasks. It excels at mathematics, science, programming, and visual reasoning, and can also be used for technical writing and multi-step analysis that follows instructions. (openai.com)
Capabilities and uses
- Accepts text and image inputs and outputs text; can reason using images and text together, making it suitable for analyzing charts, code, and complex problems.
- Can use tools in ChatGPT; supports function calling in the API, enabling the construction of multi-step workflows. The API also supports structured outputs and streaming.
- The context window is 200,000 tokens, and the maximum output is 100,000 tokens; the knowledge cutoff date is June 1, 2024. (openai.com)
Limitations and status
- The model may produce factual errors or hallucinations; important results should be independently verified.
- The API model does not support audio or video input, or fine-tuning.
- As of November 18, 2026,
o3and the snapshoto3-2025-04-16have been officially marked as deprecated, with their API service scheduled to end on December 11, 2026. (cdn.openai.com)
API pricing
The following are OpenAI's official API prices for text tokens, in U.S. dollars (USD), with each item billed per 1 million tokens; date checked: November 18, 2026.
| Billing method | Input | Cached input | Output |
|---|---|---|---|
| Standard API | $2.00 | $0.50 | $8.00 |
| Batch API | $1.00 | $0.25 | $4.00 |
Cached pricing applies only to inputs eligible for caching; Batch pricing applies to batch requests. Tool calls may incur additional charges. These prices do not represent ChatGPT subscription prices or charges on other platforms. Official pricing source: o3 model documentation. (developers.openai.com)
Release and official sources
- Release date: April 16, 2025; version snapshot:
o3-2025-04-16. - Official release announcement
- Official API model documentation and pricing
- Official deprecation schedule (openai.com)