OpenAI o3-pro

o3-pro was released on 10 June 2025, on the same day that OpenAI cut the price of ordinary o3 by 80 per cent. The two moves belong together: the company made its frontier reasoning model a commodity and, in the same breath, put a more expensive tier above it. That tier had become dramatically cheaper. o1-pro, fifteen months earlier, listed at 150 and 600 dollars per million input and output tokens. o3-pro listed at 20 and 80 — a drop of 87 per cent for the same product category. It remained four times the launch price of o3 and ten times its reduced price, so the premium over the standard model was real; what changed was the absolute number. The model is Responses API only. OpenAI warns in its own documentation that some requests take several minutes and recommends background mode to avoid timeouts — an unusually candid admission about what more thinking actually costs in wall-clock time. Its tool list is wider than o1-pro's: web search and image generation join function calling, file search and MCP. The context window is 200,000 tokens with up to 100,000 output, input is text and images, and the knowledge cutoff is June 2024. The o3-pro-2025-06-10 snapshot shuts down on 11 December 2026, together with o3 and the original GPT-5 snapshots. The replacement is gpt-5.6-sol with reasoning mode set to pro.

▸Videos

No videos yet.

▸Reviews