OpenAI Ships GPT-Image 2.5, Halving Edit Latency as Higgsfield, Adobe and Manus Put It to Work
OpenAI has begun rolling out GPT-Image 2.5, a new generation of its image model that the company says cuts editing latency by up to 50 percent while making narrow, single-instruction edits dramatically more reliable than before. The update, announced September 8, 2026 as ChatGPT Images 2.5, arrives five months after the April debut of Images 2 and reaches ChatGPT's free, Work and Codex tiers at the same time as two API-only variants: GPT-Image-2.5 Flare, the default model tuned for speed and high-volume generation, and GPT-Image-2.5 Sunburst, a slower model built for the tighter control that campaign and product-photo work demands.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.