OpenAI has dramatically reduced the cost of its advanced reasoning model, o3, making it significantly more affordable for developers while maintaining the same high-quality performance.

## Major Price Reduction

The company announced on Wednesday that it has cut prices by 80%, bringing the input cost down to just $2 per million tokens and output pricing to $8 per million tokens. This substantial reduction was achieved through optimization of the inference stack that serves the model, rather than any changes to the model itself.

“We optimized our inference stack that serves o3. Same exact model—just cheaper,” OpenAI confirmed in a social media post.

## Performance Remains Unchanged

Independent verification from ARC Prize, a benchmark community, confirmed that the o3-2025-04-16 model’s performance remained identical after the price reduction. Their testing showed no difference in results compared to the original model, proving that OpenAI achieved cost savings through technical optimization rather than model downgrades.

## Benefits for Developers

While individual ChatGPT users typically don’t interact with the API directly, this price reduction has significant implications for developers and businesses. Popular development tools that rely on the o3 API, such as Cursor and Windsurf, can now operate much more cost-effectively, potentially passing savings on to their users.

## New Premium Option

Alongside the price reduction, OpenAI has introduced the o3-pro model to its API offerings. This premium version utilizes additional computational resources to deliver enhanced results for users requiring the highest level of performance.

The combination of reduced pricing for the standard o3 model and the introduction of a premium tier demonstrates OpenAI’s strategy to make advanced AI capabilities more accessible while still catering to users with demanding performance requirements.