Editorial illustration for OpenAI upgrades ChatGPT image model, improves English text rendering
ChatGPT Image Model Gets Text Rendering Upgrade
OpenAI upgrades ChatGPT image model, improves English text rendering
Two years back, ChatGPT botched a basic task: reading a sign. That specific failure is now in the crosshairs. OpenAI's latest upgrade to its image model, dubbed Images 2.0, directly attacks its own history of garbled text.
The result? Far fewer malformed letters and phantom characters in generated images. Not to be outflanked, Google is charging the same hill, with text rendering a key target in its recent Nano Banana updates.
First Impressions After a few hours of generating images with the new model, I was generally impressed with the text rendering capabilities, in English at least. Not that long ago, image outputs featuring text, from any of the major models, often included numerous malformed characters or words with errant extra letters. ChatGPT struggled to label images accurately two years prior, so the cleaner, more complex outputs from Images 2.0 are a sign of continued improvement. Google has also focused on improving image outputs featuring text in its recent iterations of Nano Banana.
The win is immediate: you can finally get a legible storefront sign or a clean logo. That’s the pivot. It moves AI image generation from mere spectacle toward practical tool.
But OpenAI’s own release notes concede lingering stumbles with non-Latin scripts. The parallel push from Google confirms a broader industry truth. The race isn't just for better pictures anymore.
It's for utility.
Related Reading
Browse all OpenAI news →
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
Related Topic OpenAI hires 630 ex-Meta staff as ChatGPT memory may turn data into ads
Business & Startups
Related Topic Sen. Warren urges Trump admin to expand AI credit after OpenAI letter
Business & Startups
Same Week 91% of businesses now use video marketing — AI cut the cost of keeping up by 91% too
Market Trends
From Archives Starbucks ChatGPT app forces users to pick from suggested iced‑coffee options
From Archives OpenCode now supports Qwen3-Coder via config.json on Linux, macOS, Windows
From Archives OpenAI expands Trusted Access for Cyber Defense with GPT-5.4‑Cyber model
From Archives Claude targets design stack as OpenAI rebrands as major platform
Business & Startups
Common Questions Answered
How has OpenAI improved text rendering in ChatGPT image generation?
OpenAI has significantly enhanced the ability to generate images with clear, accurate text, addressing previous issues with malformed characters and extra letters. The new model (Images 2.0) now produces much cleaner text outputs, particularly in English, marking a substantial improvement over earlier versions that struggled with text accuracy.
What new capabilities does ChatGPT Images 2.0 offer for multi-image generation?
ChatGPT Images 2.0 can now generate entire study booklets from a single prompt, demonstrating enhanced multi-image creation capabilities. The system also claims support for non-English scripts like Chinese and Hindi, expanding its global usability and text rendering potential.
How is access to the new ChatGPT image generation model structured?
The new image generation service is currently live for all ChatGPT and Codex users, with a more advanced tier reserved for paying subscribers. This tiered approach allows broader access while providing additional features for premium users.
Further Reading
- The new ChatGPT Images is here — OpenAI
- OpenAI's New Image Generator Can Do Near-Perfect Text — Futurism
- OpenAI Launches GPT-Image Model 1.5 with Faster Speed and Better Text Rendering — Article Factory
- ChatGPT image generation: What's changed and why it matters — National Centre for AI
- Text That Works: GPT-4o's Conversational Approach to Image Generation — Sidecar
