History
o3-pro puts more emphasis on reliability than speed
OpenAI released a version of o3 that could spend longer working on difficult answers.

On 10 June 2025, OpenAI introduced o3-pro for Pro users in ChatGPT and through the API. The maker described it as a version of o3 that thought for longer to provide more reliable answers. OpenAI’s release notes also record an eighty per cent price reduction for o3. The model notes describe the o3-pro release.

In our view, this moment makes two separate choices visible: how much an answer costs and how much time it may take. The price reduction is historical news, rather than a current quote. Greater reliability was the maker’s aim; its meaning for your task requires separate assessment.

As a trial, a small or medium-sized business can contrast finding differences between long product lists with a simple sorting task. Define the correct results and acceptable waiting time beforehand. Use the same practice data for each comparison. Check whether extra reasoning also reduces review effort. Examine incorrect answers with convincing explanations separately. Use the results to choose a suitable route for each task. This gives the choice between quick and thorough a concrete meaning for the people doing the work.