o3-pro and the 80% o3 price cut
OpenAI released o3-pro, an o3 variant that spends more inference compute for reliability, priced at $20 / $80 per million input / output tokens (price listing; release date 2025-06-10 there, 06-11 in…
- Date
- 10 June 2025
- Who
- OpenAI
- Confidence
- Medium (prices via secondary coverage; OpenAI's pages return HTTP 403)
- Deep dive
- Reasoning II, from o1 and o3 to DeepSeek-R1 and the labs that replicated them
Tier: Supporting · Significance: 2/5 · Org(s): OpenAI · Confidence: Medium (prices via secondary coverage; OpenAI's pages return HTTP 403) OpenAI released o3-pro, an o3 variant that spends more inference compute for reliability, priced at $20 / $80 per million input / output tokens (price listing; release date 2025-06-10 there, 06-11 in other coverage), and, on 2025-06-10, cut standard o3 from $10 / $40 to $2 / $8 per million tokens, an 80% reduction (CometAPI summary, Roboin). It is the cleanest example of the two-sided pricing in B08-06, where the same lineage sold at $2 / $8 for ordinary use and at ten times that for "think longer" mode. It came 271 days after o1-preview and 141 after R1, whose price floor is the likely reason for the cut (inference). The "pro" mode line began with o1 pro.