OpenAI launched Images 2.0 with reasoning integration, multi-image generation, accurate text rendering, and a December 2025 knowledge cutoff — globally available now.
OpenAI released ChatGPT Images 2.0, a new image generation model that integrates with ChatGPT's reasoning capabilities to generate multiple images from a single prompt. The model can search the internet for real-time information, render accurate text in multiple languages including Chinese and Hindi, and has a knowledge cutoff of December 2025. It is available globally for ChatGPT and Codex users, with an enhanced tier for paid subscribers.
Images 2.0 exposes reasoning-backed image generation through the existing ChatGPT/Codex interface, meaning multi-step image workflows that previously required orchestration layers (prompt → search → generate → assemble) can now collapse into a single call. Accurate text rendering in non-English languages also eliminates a notorious post-processing headache. Check whether the multi-image output is accessible via the Codex API — if so, this could replace brittle multi-model pipelines.
If you're running a pipeline that combines a search API + image generation for infographics or data visuals, test Images 2.0 against your current setup this week — measure output accuracy and round-trip latency to see if you can drop two API calls.
Open ChatGPT at chat.openai.com and select the Images 2.0 model
Tags
Related