OpenAI's o3 model gets a massive price cut to compete with Google's AI model

Just ahead of its o3-pro rollout, the company has dramatically cut the price of its o3 model by nearly 80%. o3 delivered top-tier performance on AI benchmarks like Codeforces, SWE-bench, and MMMU.

It also came with built-in tool support for web browsing, Python, and more. But despite its capabilities, the model saw slower developer adoption. The reason? Price.

With Google’s Gemini 2.5 Pro offering similar results at a fraction of the cost—and Claude 3.7 edging it out in some coding tests—OpenAI’s o3 was facing real competition. But that’s a thing from the past now.

OpenAI slashes price of the o3 model to compete with Gemini 2.5 Pro

OpenAI has dropped o3’s pricing from $10 to $2 per million input tokens and from $40 to $8 per million output tokens. It’s a move aimed at regaining developer interest, and it could shift the balance in the ongoing model wars.

“Excited to see what people will do with it now,” wrote CEO Sam Altman on X. He also hinted that the newly announced o3-pro will offer even more performance for a price that “you’ll also be happy with.”

OpenAI’s o3-pro model, launching later today, is positioned as the company’s new flagship for reasoning. Internally, it’s expected to beat Gemini 2.5 Pro’s June update on several benchmarks. It’s worth noting that o3-pro has landed at $20 per million input tokens and $80 per million output tokens.