OpenAI's most powerful reasoning model
OpenAI o3 is a powerful reasoning AI model in the o-series, successor to o1, designed for complex tasks in coding, math, science, and visual reasoning. It uses reinforcement learning for private chain-of-thought reasoning, achieving state-of-the-art benchmarks like Codeforces Elo 2727, SWE-bench 71.7%, GPQA Diamond 87.7%. Released April 16, 2025, with variants like o3-pro and o3-mini; available via ChatGPT and API for developers, researchers, and enterprises needing advanced analysis.
Standard tier on the public pricing page; separate Batch, Flex, and Priority pricing is also published.
OpenAI’s public pricing page lists o3 under flagship models with standard API pricing of $2.00 input, $0.50 cached input, and $8.00 output per 1M tokens. The same page also lists separate Batch, Flex, and Priority rates for o3, but the standard tier is the default public price.
