OpenAI releases ChatGPT Images 2.0 with web search and multi-image generation
The updated image generator now enables web search within prompts, generates up to eight consistent images at once, and improves text rendering across multiple languages.
1 source · cross-referenced
- OpenAI has released ChatGPT Images 2.0, powered by a new GPT Image 2 model with 'thinking capabilities' that allow web search integration during image generation.
- The thinking feature, available to ChatGPT Plus, Pro, Business, and Enterprise subscribers, enables the tool to generate up to eight images simultaneously while maintaining consistent characters and styles.
- New capabilities include higher resolution output (up to 2K), expanded aspect ratios (3:1 to 1:3), and significant improvements in rendering text in non-Latin scripts including Japanese, Korean, Chinese, Hindi, and Bengali.
- All ChatGPT users gain access to improvements in photo detail capture, pixel art, manga, and cinematic image generation, plus faster generation speeds and enhanced photo editing.
OpenAI has rolled out ChatGPT Images 2.0, introducing AI-powered image generation with integrated web search capabilities. The system operates via a new GPT Image 2 model and uses what OpenAI calls 'thinking capabilities'—a mechanism that allows the tool to reason through image structure and pull information from the web before generating outputs.
For ChatGPT's paid tiers (Plus, Pro, Business, and Enterprise), the thinking mode unlocks multi-image generation: users can now create up to eight images in a single batch while preserving consistent characters, objects, and visual styles across the set. The feature targets use cases like manga page sequences, design variations across multiple rooms, or cohesive social media graphics.
Technical improvements span all ChatGPT user tiers. The tool now generates images at resolutions up to 2K and supports wider aspect ratios—from ultra-wide (3:1) to portrait-oriented (1:3) formats. Language rendering has expanded beyond Latin scripts, with OpenAI claiming 'significant gains' in text generation for Japanese, Korean, Chinese, Hindi, and Bengali.
The update arrives as image generation competition intensifies. Google and Microsoft have both released competing tools in recent months, putting pressure on OpenAI to expand feature depth and usability. ChatGPT Images 2.0 is available to all ChatGPT and Codex users beginning on the announcement date.
- Apr 24, 2026 · TechCrunch
Sierra acquires YC-backed AI workflow startup Fragment
Trust54 - Apr 23, 2026 · Hugging Face
Gemma 4 Vision-Language Model Demo Runs on Edge Device With Local Audio and Webcam
Trust70 - Apr 23, 2026 · OpenAI — News
OpenAI describes WebSocket optimization for agent API performance
Trust79