--- title: OpenAI releases Images 2.0 with better text rendering... description: OpenAI released two major updates: GPT Image 2 (gpt-image-2) and GPT-5.5. GPT Image 2 features dramatically improved text rendering (~99% accuracy) across... canonical: https://daily.dev/posts/openai-releases-images-2-0-with-better-text-rendering-and-a-thinking-mode-qimpeo76v twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: OpenAI releases Images 2.0 with better text rendering and a thinking mode | daily.dev og:description: OpenAI released two major updates: GPT Image 2 (gpt-image-2) and GPT-5.5. GPT Image 2 features dramatically improved text rendering (~99% accuracy) across... og:url: https://daily.dev/posts/openai-releases-images-2-0-with-better-text-rendering-and-a-thinking-mode-qimpeo76v og:image: https://api.daily.dev/og/posts/qImpeO76V.png og:image:alt: OpenAI releases Images 2.0 with better text rendering and a thinking mode og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # OpenAI releases Images 2.0 with better text rendering and a thinking mode **[Collections](https://daily.dev/sources/collections)** · 3 min read · 0 upvotes · 0 comments ## Summary OpenAI released two major updates: GPT Image 2 (gpt-image-2) and GPT-5.5. GPT Image 2 features dramatically improved text rendering (~99% accuracy) across non-Latin scripts, two generation modes (Instant and Thinking), up to 8 images per prompt, and resolutions up to 2K-4K. It topped the Image Arena leaderboard within 12 hours of launch, displacing Google's Gemini. It's available to all ChatGPT users and via API, with Vercel AI Gateway integration. DALL-E 2 and 3 will be deprecated in May 2026. GPT-5.5 is optimized for agentic coding, computer use, and scientific research, scoring 58.6% on SWE-Bench Pro and 85.0% on ARC-AGI-2. It uses 50-60% fewer tokens on coding tasks and is priced at $5/M input and $30/M output tokens. It's rolling out to Plus, Pro, Business, and Enterprise users. ## Content ## GPT Image 2 OpenAI launched ChatGPT Images 2.0 (also called gpt-image-2) with what's probably its most meaningful image generation upgrade in years. The headline improvement is text rendering — historically a disaster for AI image models — which now works reliably across non-Latin scripts including Japanese, Korean, Hindi, and Bengali. OpenAI claims roughly 99% text rendering accuracy. The model runs in two modes: **Instant** for fast output and **Thinking** for deliberate, multi-step generation. In Thinking mode, the model plans composition, reasons about spatial relationships, and can search the web for real-time context before drawing anything. It can also generate up to 8 coherent images from a single prompt and self-check its own outputs. Resolution goes up to 2K (with some reports of 4K being available), and it supports flexible aspect ratios for things like marketing assets and multi-panel comics. Within 12 hours of launch, gpt-image-2 took the top spot on the Image Arena leaderboard by a record margin — displacing Google's Gemini, which had been leading. Simon Willison tested it against gpt-image-1 and Google's models using a complex Where's Waldo-style prompt (a raccoon with a ham radio in a detailed crowd scene). At high quality and 3840x2160 resolution, gpt-image-2 produced a clearly superior result, though at around 40 cents per image. Practical use cases people are already running with: generating open graph images from URLs, extracting and fixing perspective on phone screenshots, creating transparent cutouts, redesigning landing page hero sections, and building mood boards for app design. The GPT-5.5 + Image 2 combination seems to be a popular workflow — describe a design language in GPT-5.5, iterate on it, then generate visuals. The model is available to all ChatGPT users and via API as `gpt-image-2`. It's also live on Vercel's AI Gateway, which means developers can access it without a separate OpenAI account or markup fees. Figma integration is rolling out too. The Thinking mode and multi-image features are paywalled behind Plus/Pro tiers. DALL-E 2 and DALL-E 3 are being deprecated on May 12, 2026. OpenAI hasn't disclosed the underlying architecture. Adoption so far has been uneven globally. India had roughly 5 million downloads during launch week versus about 2 million in the US, with strong spikes in Pakistan, Vietnam, and Indonesia. The improved Hindi and Bengali text rendering is likely a factor. Overall daily active users were up about 1% week-over-week globally — solid but not explosive outside emerging markets. --- ## GPT-5.5 OpenAI also released GPT-5.5, positioned as its most capable model to date and optimized for agentic coding, computer use, and scientific research. Benchmark numbers: 82.7% on Terminal-Bench 2.0, 58.6% on SWE-Bench Pro, 78.7% on OSWorld-Verified, and 85.0% on ARC-AGI-2. A GitHub VP noted it uses 50-60% fewer tokens on coding tasks compared to previous models, which matters a lot for cost at scale. GPT-5.5 Pro achieved a new high score of 159 on the Epoch Capabilities Index. Latency is roughly on par with GPT-5.4 per token, so the efficiency gains come from doing more with fewer tokens rather than raw speed. Pricing is $5 per million input tokens and $30 per million output tokens. The model was co-designed for NVIDIA GB200/GB300 NVL72 systems. It's rolling out to Plus, Pro, Business, and Enterprise users in ChatGPT and Codex, with API access coming soon. A GPT-5.5 Pro variant is available for higher-accuracy work. Sam Altman's framing on the release: iterative deployment is part of OpenAI's safety strategy, the goal is democratization of access, and rapid improvements are expected. The model includes updated cybersecurity and biosafety safeguards under OpenAI's Preparedness Framework. ## Similar posts on daily.dev - [OpenAI launches GPT-5.5, calling it “a new class of intelligence”](https://daily.dev/posts/openai-launches-gpt-5-5-calling-it-a-new-class-of-intelligence--tqwziccn1) · The New Stack · 1 upvotes · 1 comments - [I've been using GPT Image 2 all week — here's what it's actually good for](https://daily.dev/posts/i-ve-been-using-gpt-image-2-all-week-here-s-what-it-s-actually-good-for-fvc6rn4re) · Visually AI · 1 upvotes · 0 comments --- Tags: [#genai](https://daily.dev/tags/genai), [#openai](https://daily.dev/tags/openai), [#chatgpt](https://daily.dev/tags/chatgpt) [View this post on daily.dev](https://daily.dev/posts/openai-releases-images-2-0-with-better-text-rendering-and-a-thinking-mode-qimpeo76v) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"OpenAI releases Images 2.0 with better text rendering and a thinking mode","url":"https://daily.dev/posts/openai-releases-images-2-0-with-better-text-rendering-and-a-thinking-mode-qimpeo76v","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/openai-releases-images-2-0-with-better-text-rendering-and-a-thinking-mode-qimpeo76v"},"datePublished":"2026-04-21T19:23:01.092Z","dateModified":"2026-05-01T02:03:29.894Z","description":"OpenAI released two major updates: GPT Image 2 (gpt-image-2) and GPT-5.5. GPT Image 2 features dramatically improved text rendering (~99% accuracy) across...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/c10bdaaafae488b8d4fee9a7809b7102?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/c10bdaaafae488b8d4fee9a7809b7102?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Collections","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Collections","logo":"https://media.daily.dev/image/upload/s--fk_6ycEi--/f_auto,q_auto/v1780996001/logos/collections?_a=BAMAMiWQ0","url":"https://daily.dev/sources/collections"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/openai-releases-images-2-0-with-better-text-rendering-and-a-thinking-mode-qimpeo76v","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"genai,openai,chatgpt","timeRequired":"PT3M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Collections","item":"https://daily.dev/sources/collections"},{"@type":"ListItem","position":3,"name":"OpenAI releases Images 2.0 with better text rendering and a thinking mode"}]} ```