GPT Image 1 mini

GPT Image 1 mini was released on 6 October 2025 at OpenAI DevDay, in the same wave as GPT-5 Pro, Sora 2 and the mini variants of the realtime and audio models. It is a cost-efficient version of GPT Image 1: a natively multimodal model that takes text and images and returns images. Price is the whole argument for it. A low-quality 1024x1024 image costs $0.005, a medium one $0.011 and a high-quality one $0.036 — roughly a quarter of what the same picture costs from GPT Image 1.5 at every tier. Image tokens are $2.5 per million on input and $8 on output, text input $2 per million, and cached input drops to a tenth of the list rate, which matters for editing workflows that resend the same reference image. On 2 June 2026 OpenAI put it on the shutdown list for 1 December 2026, alongside GPT Image 1.5, and named GPT Image 2 as the replacement — a model priced well above it, which leaves the cheap tier of OpenAI image generation without a direct successor.

▸Videos

No videos yet.

▸Reviews