OpenAI Cuts O3 Pricing by 80% as a Bold Move Against Gemini 2.5 Pro

OpenAI’s Game-Changing Price Reduction for o3 Model

In April, OpenAI launched its leading reasoning model, o3, which achieved remarkable success in various AI benchmarks, including Codeforces, SWE-bench, and MMMU. This model not only set a new standard for performance but also provided capabilities for utilizing external tools like web browsing and a Python interpreter.

Market Response and Competitive Landscape

Despite its advanced features, the o3 model wasn’t widely embraced by developers primarily due to its higher cost compared to competitors. For instance, Google’s Gemini 2.5 Pro offered similar performance at only a quarter of the price, while Anthropic’s Claude 3.7 exhibited superior coding abilities at comparable rates.

Significant Price Reduction Announced

In a strategic move to enhance adoption, OpenAI has dramatically reduced the o3 model’s pricing by nearly 80%.Initially costing $10 per million input tokens and $40 per million output tokens, the new rates are now set at $2 per million input tokens and $8 per million output tokens. This substantial price cut positions o3 as a more appealing choice for developers looking for high-performance AI solutions.

we dropped the price of o3 by 80%!! excited to see what people will do with it now.think you’ll also be happy with o3-pro pricing for the performance:)

— Sam Altman (@sama) June 10, 2025

Benchmark Comparisons Revealed

To better understand how OpenAI’s o3 stacks against other leading models, refer to the benchmark comparison chart shared by Google during the launch of the Gemini 2.5 Pro June update. This detailed analysis highlights strengths and weaknesses across various AI systems.

Introduction of o3 Pro Model

In addition to the price adjustments, OpenAI is set to unveil the o3 Pro model today. This advanced version is anticipated to exceed the performance of the Gemini 2.5 Pro June update across key AI benchmarks and may be priced at $10 per million input tokens and $40 per million output tokens.

Source & Images