OpenAI o3-mini

o3-mini went live on 31 January 2025 as the small reasoning model of the o3 generation. OpenAI described it plainly: the intelligence of a larger reasoning model at the cost and latency targets of o1-mini, with developer features the earlier small model lacked — structured outputs, function calling and Batch API support. It reads and writes text only, with a 200,000-token context window, up to 100,000 output tokens and support for reasoning tokens, which are billed as output. The knowledge cutoff is 1 October 2023, the oldest of any model in this part of the OpenAI line-up. Pricing is $1.1 per million input tokens, $0.55 cached and $4.4 per million output. On 22 April 2026 OpenAI scheduled it for shutdown on 23 October 2026, with GPT-5.6 Sol named as the replacement. Its own successor, o4-mini, was put on the same list with the same date, which retires both small reasoning models of that era on a single day.

▸News

Eleven OpenAI models switch off on one day — and three GPT-5.6 variants inherit all of them

8/20/2026 · OpenAI — Deprecations

▸Videos

No videos yet.

▸Reviews