OpenAI Launches ChatGPT Images 2.0 With Thinking Level Intelligence for Design

OpenAIOpenAI

· Updated

OpenAI released ChatGPT Images 2.0, a new image model capable of complex visual reasoning and precise multilingual text rendering. This update shifts image generation from simple art creation to functional design, allowing users to produce publication-ready infographics and layouts directly.

OpenAI launched ChatGPT Images 2.0, which extends spatial reasoning to handle complex visual tasks and precise editing. Powered by gpt-image-2, it uses "thinking-level intelligence" to plan multi-step design compositions and data visualizations with an updated knowledge cutoff of December 2025.
Underlying model
gpt-image-2
Knowledge cutoff
December 2025
Thinking mode availability
Plus, Pro, and Business tiers
API access
gpt-image-2 available now
Supported scripts
Latin, Arabic, Devanagari, Japanese, Cyrillic

This release builds on the reasoning paradigm seen in GPT-5.4 by applying it to the visual domain. Moving beyond simple prompt-to-image generation, the model handles end-to-end workflows like creating academic posters or magazine spreads. It also accurately renders text in scripts like Arabic, Devanagari, and Japanese.

Access is live today across all ChatGPT and Codex tiers, though the reasoning-enhanced "thinking" mode is restricted to Plus, Pro, and Business users. Developers can integrate these capabilities via the gpt-image-2 API, which supports flexible aspect ratios and high-fidelity product photography.

View the full update on openai.com
OpenAI
OpenAI
@OpenAI
X

Introducing ChatGPT Images 2.0 A state-of-the-art image model that can take on complex visual tasks and produce precise, immediately usable visuals, with sharper editing, richer layouts, and thinking-level intelligence. Video made with ChatGPT Images https://t.co/3aWfXakrcR

676retweets6.9klikes
View on X

Still wondering? A few quick answers below.

It is a new image generation model from OpenAI that handles complex visual tasks like layout design and precise text rendering. It uses thinking-level intelligence to reason through prompts, allowing it to create functional assets like infographics and posters rather than just simple artistic images.

The model is available to all ChatGPT and Codex users starting today. However, the advanced thinking capabilities are restricted to ChatGPT Plus, Pro, and Business subscribers. Enterprise access is coming soon. Developers can also access the underlying gpt-image-2 model through the OpenAI API for custom application integration.

Yes, the model features significantly improved multilingual text rendering. It can accurately generate text in various scripts, including Japanese, Arabic, Devanagari, Cyrillic, Bengali, and Greek. This allows users to create localized marketing materials, manga, and editorial content with precise typography across global languages and scripts.

ChatGPT Images 2.0 has an updated knowledge cutoff of December 2025. This real-world intelligence allows the model to understand and visualize more recent events, products, and cultural contexts compared to previous versions. It can handle tasks end-to-end, from initial copywriting and research to final design composition and layout.

Unlike earlier models that focused primarily on visual fidelity, ChatGPT Images 2.0 introduces thinking-level intelligence. This allows the model to perform advanced visual reasoning, such as following complex instructions for academic posters, mathematical proofs, and multi-panel comic sequences. It acts as a visual thought partner that can research and transform source materials into polished assets.

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

OpenAI

See all AI news & updates from OpenAI →

Share this update