Google Unveils Gemini 2.5 Pro Preview (I/O Edition) with Enhanced Coding Capabilities

Google debuts an updated Gemini 2.5 Pro AI model ahead of I/O | TechCrunch

Google on Tuesday announced the launch of Gemini 2.5 Pro Preview (I/O edition), an updated version of its flagship Gemini 2.5 Pro AI model that the company claims tops a number of widely-used benchmarks. Gemini 2.5 Pro Preview (I/O edition) is available via the Gemini API and Google's Vertex AI and AI Studio platforms, priced the same as the Gemini 2.5 Pro model it effectively replaces. It's also in Google's Gemini chatbot app for the web and mobile devices. The model's release comes ahead of Google's annual I/O developer conference (hence the "I/O edition" designation), where Google is expected to unveil a host of models as well as AI-powered tools and platforms. The company is competing fiercely for mindshare and marketshare in the cutthroat AI race; rivals like OpenAI and xAI are on the cusp of releasing models that are expected to be highly performant. According to Google, Gemini 2.5 Pro Preview (I/O edition) has "significantly" improved capabilities for coding and building interactive web apps. The model is also better at tasks like code transformation -- i.e. modifying a piece of code to achieve a specific goal -- and code editing, the company says. In a blog post, Google notes that Gemini 2.5 Pro Preview (I/O edition) leads the WebDev Arena Leaderboard, a benchmark measuring a model's ability to create aesthetically pleasing and functional web apps. Gemini 2.5 Pro Preview (I/O edition) also has state-of-the-art performance in video understanding, achieving a score of 84.8% on one popular benchmark, VideoMME. "For developers already using Gemini 2.5 Pro, this new version will not only improve coding performance but will also address key developer feedback including reducing errors in function calling and improving function calling trigger rates," wrote Google in a blog post. "By default, the model has a real taste for aesthetic web development while maintaining its steerability [...] [And] with improved video understanding and complete UI, the updated Gemini 2.5 Pro model delivers a more functional experience than the previous simple example."

Gemini 2.5 Pro Update Boosts Coding Abilities Ahead of Google I/O Splash

Imad is a senior reporter covering Google and internet culture. Hailing from Texas, Imad started his journalism career in 2013 and has amassed bylines with The New York Times, The Washington Post, ESPN, Tom's Guide and Wired, among others. Google is releasing an updated version of Gemini 2.5 Pro ahead of its Google I/O conference later this month that it says has significantly improved coding capabilities. Gemini 2.5 Pro Preview (I/O Edition) is especially good at "building compelling interactive web apps," the company said in a press release on Tuesday. It's an area with intense competition among Google and its AI rivals. In February, Google released Gemini's free Code Assistant, a model that gave users "practically unlimited" use at 180,000 completions per month. But to get the maximum out of it, users would have had to code for 14 hours a day, every day. The coding assistant in Gemini 2.5 Pro is a more powerful option, and Google says Gemini 2.5 Pro now leads on the WebDev Arena Leaderboard, surpassing its previous version. Google, Anthropic, Meta and ChatGPT maker OpenAI have all been trading blows to be No. 1 on benchmarking leaderboards. For companies, it's a good indicator of who's dominating in that arena and fuels future funding and development. There's been criticism, though, that benchmarking tools aren't indicative of actual real-world effectiveness and that Big Tech companies have been tuning AI models for benchmarks rather than for how actual people use AI. Some experts are calling for AI companies to use different standards when benchmarking, not focusing so much on graduate-level analysis or abstract math, but rather parameters that test real-world needs. Still, the leaderboard competition does give consumers and enterprise customers some indication of how various AI models stack up. "We're excited about the latest Gemini 2.5 Pro, which builds on its already strong real-world coding capabilities," said Tulsee Doshi, Google's senior director of product management for Gemini, in a statement. "We're observing internally that the new model has a significant reduction in its failure to call tools, an improvement we believe our users will find makes 2.5 Pro even more effective than before in Cursor." Cursor is an AI-powered coding environment for computers. Google says it has received "overwhelmingly positive feedback" for Gemini 2.5 Pro. Users on Reddit, meanwhile, report that Gemini 2.5 Pro still has a mind of its own and doesn't follow instruction.

Google's Gemini 2.5 Pro update makes the AI model even better at coding

The model update was scheduled to debut at Google I/O, but Google released it early in response to positive feedback from users. Google I/O, the company's annual developer event where it unveils its latest and greatest features, is only a couple of weeks away. Google is offering a glimpse of I/O announcements with its latest AI model launch. On Tuesday, Google released early access to Gemini 2.5 Pro Preview (I/O edition), a version of Gemini 2.5 Pro that has significantly improved coding capabilities -- especially when building interactive web apps, according to the release. The update is also meant to make the model better at other coding tasks, such as code transformation, code editing, and creating agentic workflows. Also: Why I just added Gemini 2.5 Pro to the very short list of AI tools I pay for The model update was scheduled to debut at Google I/O, but Google released it early in response to positive feedback so users could start building as soon as possible. The updates place Gemini 2.5 Pro at the top of the WebDev Arena Leaderboard, with an Elo score 147 points higher than its predecessor. It also scored 84.8% on the VideoMME benchmark. The updated version of Gemini 2.5 Pro is available in the Gemini API for developers via Google AI Studio and Vertex AI. Users in the general public can also access the updated model in the Gemini app, where it powers features like Canvas. Also: Gemini for kids is rolling out, and it comes with 3 warnings from Google Google I/O will take place on May 20 to May 21 at the Shoreline Amphitheatre in Mountain View, California. The keynote event, which will feature the bulk of the news regarding Google's latest hardware and software, will take place on Day One at 10 a.m. PT / 1 p.m. ET. Also: Google's AI Mode may be the upgrade Search desperately needs - how to try it for free Developers can register for the event's digital experience on the Google I/O landing page now for free. Everyone is invited to tune into the live stream at no cost. ZDNET will also cover the event on the ground with updates posted to the site as the news happens. Get the morning's top stories in your inbox each day with our Tech Today newsletter.

Google updates Gemini 2.5 Pro model for coders

Google has updated its Gemini 2.5 Pro AI model with stronger coding capabilities, hoping developers would begin building with the model before the Google I/O developer conference later this month. Released May 6, Gemini 2.5 Pro Preview (I/O edition) is accessible in Google AI Studio tool and in Vertex AI. Commenting on the update, Google said developers could expect meaningful improvements for front-end and UI development alongside improvements in fundamental coding tasks including transforming and editing code and building sophisticated agentic workflows. Developers already using Gemini 2.5 Pro will find not only improved coding performance but reduced errors in function calling and improved function calling trigger rates, the company said.

Google has a special edition of Gemini Pro for I/O 2025

Summary Google will showcase AI and Android advancements at Google I/O 2025. Gemini 2.5 Pro Preview (I/O edition) is now available, making coding simpler, including front-end and UI development. The new Gemini model reduces coding errors and is available for free to current users. We're just a couple of weeks out from Google I/O 2025, set to take place on May 20 and 21. There's been a lot of talk about what could be shown at the event, but if the past is any indication, we know that the brand will be focusing its core efforts on showing off what's next with Android, while also showcasing just how far it's gotten with its AI efforts. Related Gemini leans into coding with new GitHub integration and a web tool Let AI navigate repositories for you Posts And to give developers a taste, the brand has now released Gemini 2.5 Pro Preview (I/O edition), which will make it even easier to bring code to life, with new changes, like "front-end and UI development, alongside improvements in fundamental coding tasks such as transforming and editing code." Starting things early this year Since the introduction of AI, we've seen a number of changes sweep through many different industries, making mundane tasks easy to accomplish with a little bit of assistance. Google has been making huge strides with Gemini, bringing the tech to a variety of different products in its portfolio, allowing users to see just how versatile it can be. Of course, coding is one area where AI really excels, which makes, according to Google, "Gemini 2.5 Pro the go-to model for developers." Google was quick to highlight that its new model will be able to extract information from a video and create apps based on what it's learned. Gemini 2.5 Pro will also be able to make coding simpler by doing the heavy lifting when it comes to mundane tasks like changing the style and look of features on an app and matching themes when needed. You can see an example of this in the video above, where a prompt is created and Gemini easily complies to make it happen. There have also been some refinements made that were previously requested, like "reducing errors in function calling and improving function calling trigger rates." Of course, Gemini 2.5 Pro will have all kinds of uses, and these are just some examples of how it can be used today. So if you've been itching to get your hands on some new tools, give it a try. You can head into the Google AI Studio to access the updated Gemini API. Those already using Gemini 2.5 Pro should see the new features available and there isn't any extra subscription fee required for the latest model. Without a doubt, we'll be hearing more about this during Google I/O in just a couple of weeks' time.

Google announces Gemini 2.5 Pro (I/O Edition) with coding upgrades

Google is updating Gemini 2.5 Pro with "significantly improved" coding capabilities. This updated model was originally going to be released at I/O 2025, but the company decided to move up the launch. Cheekily referred to as "Gemini 2.5 Pro (I/O Edition)," Google touts improvements in "code transformation, code editing, and developing complex agentic workflows." It follows the initial (03-25) release in late March. Google is particularly touting how Gemini 2.5 Pro surpassed the previous #1 on the WebDev Arena Leaderboard by 147 Elo points. (2.5 Pro was previously at #2.) In terms of benchmarks, LiveCodeBench v5 (code generation) goes from 70.4% to 75.6%. This leaderboard measures human preference for a model's ability to build aesthetically pleasing and functional web apps. Google also touts "state-of-the-art performance in video understanding, with a score of 84.8% on the VideoMME benchmark." This update to 2.5 Pro (gemini-2.5-pro-preview-05-06) is available in the Gemini app, with Canvas particularly benefiting (as seen below), and the Gemini API (Google AI Studio + Vertex AI). Google moved up the release by two weeks given "overwhelming enthusiasm for this model."

Google just unveiled a major update to Gemini AI ahead of I/O -- here's what it can do

Google has announced its latest update to the Gemini AI model, focusing all of its energy on one particular area. To keep up with the competition from OpenAI, Gemini is now an absolute coding machine. This latest update is called Gemini 2.5 Pro Preview (I/O Edition). It's a mouthful, but surprisingly makes sense. I/O is Google's yearly conference where they announce the latest projects they are working on, and it's happening in just two weeks time. While Google will almost definitely reveal further Gemini updates at this event, this is a glimpse into the direction it is heading. According to Google, this model update preview is focused on coding and building interactive web apps. It is better at coding-based prompts across the board, but more specifically, code transformation -- the ability to modify a piece of code to fit a specific goal -- and editing existing code. Announcing the launch of this latest update in a blog post and on X, Google stated that this model tops the WebDev Arena Leaderboard. This is a benchmark that focuses on an AI model's ability to create aesthetically pleasing and functional apps. While this update won't be of much interest to the average person developers can start building using the tool in the updated Gemini 2.5 Pro in the Gemini API via Google Studio. It can also be used in the Gemini app, where it powers features like Canvas. Through this, anyone can vibe code and build interactive apps. Vibe coding is a growing trend in the AI world where users can create apps and websites without needing to know how to code, focusing on the use of prompts instead. In the world of AI, there are a few tasks that the big-name companies are trying to race to the top for. These include AI image and video generation, the ability to think through complicated tasks, and coding. Coding through AI has seen a huge uptick in ability. xAI and OpenAI are both pouring money into it, and so is Gemini. Recently, OpenAI spent a whopping $3 billion to buy Windsurf, an AI coding company. xAI's Grok also saw a major boost to its coding abilities in its latest update, focusing on the ability to help users with vibe coding. As AI becomes more accustomed to certain tasks, coding seems to be the next big barrier, and that barrier is already quickly falling.

Meet the new king of AI coding: Google's Gemini 2.5 Pro I/O Edition dethrones Claude 3.7 Sonnet

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More There's a new king on the throne of AI coding models: Today, Google's DeepMind AI research unit unveiled Gemini 2.5 Pro "I/O" edition, a new version of its hit Gemini 2.5 Pro multimodal large language model (LLM) released back in March that DeepMind CEO Demis Hassabis said on X is "the best coding model we've ever built!" Indeed, the initial benchmarks released by the company indicate Google has taken the lead -- for the first time since the generative AI race began in earnest with the late 2022 launch of ChatGPT -- above all other models on at least one important coding benchmark. The new version, labeled "gemini-2.5-pro-preview-05-06," replaces the previous 03-25 release and is now available for indie developers in Google AI Studio and for enterprises in the Vertex AI cloud platform, as well as to individual users in the Gemini app. Google's blog post said it also powers the Gemini mobile app's Canvas and other features. The new version powers feature development in apps like Gemini 95, where the model helps match visual styles across components automatically. It also enables workflows like converting YouTube videos into full-featured learning applications and crafting highly styled components -- such as responsive video players or animated dictation UIs -- with little to no manual CSS editing. It's a proprietary model, meaning enterprises will have to pay Google to use it and access it only through Google's web services. However, it doesn't alter pricing or rate limits; current users of Gemini 2.5 Pro will be automatically routed to the updated model which costs $1.25/$10 per million tokens in/out (for context lengths of 200,000 tokens) compared to Claude 3.7 Sonnet's $3/$15. The company frames this move -- ahead of Google's annual I/O (input/output) developer conference later this month in Mountain View and online, May 20-21 -- as a response to strong community feedback around Gemini's practical utility in real-world code generation and interface design. Logan Kilpatrick, Senior Product Manager for Gemini API and Google AI Studio, confirmed in a developer blog post that the update also addresses key developer feedback around function calling, with improvements in error reduction and trigger reliability. Top scores from human raters at generating web apps On WebDev Arena Leaderboard, a third-party metric that ranks models by human preference based on their ability to generate visually appealing and functional web apps, Gemini 2.5 Pro Preview (05-06) has now overtaken Anthropic's Claude 3.7 Sonnet at the number one spot. The new version scored 1499.95 on the leaderboard, placing it well ahead of Sonnet 3.7's 1377.10. The previous Gemini 2.5 Pro (03-25) model held third place with a score of 1278.96, meaning the I/O edition represents a 221-point jump. As noted by the AI power user "Lisan al Gaib" on X, not even OpenAI's GPT-4o ("o3") was able to displace Sonnet 3.7, highlighting the significance of Gemini's advancement. Gemini's performance boost reflects improved reliability, aesthetics, and usability in its outputs. Already winning rave reviews Several developers and platform leaders have highlighted the model's improved reliability and application in production scenarios. Cognition's Silas Alberti noted that Gemini 2.5 Pro was the first model to successfully complete a complex refactoring of a backend routing system, demonstrating the kind of decision-making one would expect from a senior developer. Michael Truell, CEO of the AI coding tool Cursor, said internal testing shows a marked decrease in tool call failures, a previously noted issue. He expects users to find the latest version significantly more effective in hands-on environments. Cursor has already integrated Gemini 2.5 Pro into its own code agent, reflecting how developers are using the model as a key component in more intelligent developer workflows. Michele Catasta, President of Replit, described Gemini 2.5 Pro as the best frontier model for balancing capability with latency. His comments suggest that Replit is considering integration of the model into its own tools, especially for tasks where high responsiveness and reliability are crucial. Similarly, AI educator and BlueShell private AI chatbot founder Paul Couvert noted on X that "Its code and UI generation capabilities are impressive.'" And as Pietro Schirano, CEO of the AI art tool EverArt, noted on X, the new Gemini 2.5 Pro I/O edition was able to generate an interactive simulation of the "1 gorilla vs. 100 men" meme that's been circulating on social media lately from a single prompt. Showing off another interactive Tetris-style puzzle game with working sound effects reportedly created in less than a minute, X user "RameshR" (@rezmeram) wrote that "the casual game industry is dead!!" These endorsements add weight to DeepMind's claims of practical improvements and may encourage broader adoption across developer platforms. Full apps and programs from one text prompt One of the standout features of the update is its ability to build full, interactive web apps or simulations from a single prompt. This aligns with DeepMind's vision of simplifying the prototyping and development process. Demonstrations within the Gemini app showcase how users can transform visual patterns or thematic prompts into usable code, lowering the barrier to entry for design-oriented developers and teams experimenting with new ideas. Although the architecture and under-the-hood changes of Gemini 2.5 Pro have not been detailed publicly, the emphasis remains on enabling faster, more intuitive development experiences. By leaning into its strengths in code generation and multimodal inputs, Gemini 2.5 Pro is positioned less as a research novelty and more as a practical tool for real-world coding challenges. The early release reflects a clear intention from Google DeepMind to meet developer demand and maintain momentum ahead of its major conference announcements.

Google's Gemini 2.5 Pro Now Best at Building Web Apps

The model now leads the WebDev Arena Leaderboard, which measures its capability to develop visually pleasing and functional web apps according to human preferences. It tops the chart with a score of 1419.95 points, with Anthropic's Claude 3.7 Sonnet in second place, scoring 1357.10 points. Michael Truell, CEO of Cursor, was all praise for Gemini 2.5 Pro, and said, "We're observing internally that the new model has a significant reduction in its failure to call tools, an improvement we believe our users will find makes 2.5 Pro even more effective than before in Cursor." The model is available for developers through the Gemini API in Google AI Studio and Vertex AI. Additionally, the new model is also accessible for users on the Gemini app, and its web app development capabilities can be harnessed using the built-in Canvas feature. Having said that, Google's Gemini 2.5 Pro model was already ranked highly, both on benchmarks and by developers with real-world coding experience. In the Aider Polyglot leaderboard, which evaluates LLMs' capabilities in writing and editing code, Gemini 2.5 Pro Preview scored 72.9%. It performed better than Claude 3.7 Sonnet (64.9%), OpenAI's o1 (61.7%), and the o3-mini high at 60.4%. However, OpenAI's newly released o3 model scored higher than the Gemini 2.5 Pro with a score of 79.6%. On several other benchmarks, Google's Gemini 2.5 Pro is either second to or on par with OpenAI's new o3 and o4-mini model on several other benchmarkss. Interestingly, Google said that the update was supposed to be released in the next 'couple of weeks', but the company said that, "But based on the overwhelming enthusiasm for this model, we wanted to get it in your hands sooner so people can start building." Note that this update comes only hours after reports emerged that OpenAI has reached an agreement to buy Windsurf, the AI-enabled coding platform for $3 billion. This marks OpenAI's attempt to further strengthen the coding capabilities it can offer to users -- alongside the already powerful o3 and o4-mini models, and the Codex CLI tool. To further intensify the competition, Anysphere, the company behind Cursor, the AI-enabled coding platform, recently raised a $900 million investment, led by Thrive Capital, Andreessen Horowitz (a16z), and Accel Ventures.

Google enhances coding capabilities in Gemini 2.5 Pro Preview before I/O 2025 dev conference - SiliconANGLE

Google enhances coding capabilities in Gemini 2.5 Pro Preview before I/O 2025 dev conference Google LLC today rolled out an early access release of Gemini 2.5 Pro Preview, an updated version of the company's artificial intelligence flagship model with significantly improved coding capabilities. Google said it intended to announce the Gemini 2.5 Pro Preview model at Google I/O 2025, the company's upcoming annual developer conference in a couple of weeks, but due to overwhelming positive feedback for the model, it decided to release it early. The previous version of the Gemini 2.5 Pro model, released in March built with chain-of-thought reasoning, making it the "most intelligent" model the company has released, Google said. The model is one of the most capable for coding and building user interfaces using its multimodal understanding and reasoning capabilities. "Beyond UI-focused development, these improvements extend to other coding tasks such as code transformation, code editing and developing complex agentic workflows," said Tulsee Doshi Senior, director of product management at Google. With the enhancements, Google said the model now leads in the WebDev Arena Leaderboard, exceeding the score of its previous version by 147 points. The leaderboard measures human preference for a model's ability to build web apps that are both functional and understandable. The Gemini 2.5 Pro Preview also maintains an industry large standard context window of 1 million tokens, which allows it to ingest extremely vast datasets, including big documents and codebases, around 1 hour of video or 11 hours of audio. The company has mentioned it intends to extend the window to 2 million. "We're observing internally that the new model has a significant reduction in its failure to call tools, an improvement we believe our users will find makes 2.5 Pro even more effective than before in Cursor," said Michael Truell, chief executive of Anysphere Inc., developer of the Cursor AI-powered coding editor.

Gemini Pro preview - Even better coding performance

Android Gemini Pro preview - Even better coding performance Tuesday, May 6, 2025 Richard Harris Developers gain powerful new tools for UI design, video-to-code workflows, and agentic programming with the Gemini Pro Preview - even better coding performance, now enhanced to deliver faster, more reliable coding experiences across modern web apps. Developers had been doing amazing things with Gemini 2.5 Pro, so an updated version was released a couple of weeks early to get into developers' hands sooner. Gemini 2.5 Pro Preview (I/O edition) was recently released. This update features even stronger coding capabilities, for developers to start building with before Google I/O later this month. Meaningful improvements were made for front-end and UI development, alongside improvements in fundamental coding tasks such as transforming and editing code, and creating sophisticated agentic workflows. Gemini Pro Preview: Even better coding performance "We found Gemini 2.5 Pro to be the best frontier model when it comes to 'capability over latency' ratio. I look forward to rolling it out on Replit Agent whenever a latency-sensitive task needs to be accomplished with a high degree of reliability," said Michele Catasta, President, Replit. improved coding performance on WebDev Arena Best-in-class frontend web development Gemini 2.5 Pro now ranks #1 on the WebDev Arena leaderboard, which measures human preference for a model's ability to build aesthetically pleasing and functional web apps. Drawing on this leading capability, Gemini 2.5 Pro powers Cursor's innovative code agent and empowers collaborations with companies like Cognition and Replit. Together, the frontiers of agentic programming are being pushed to unlock new possibilities for developers. "The updated Gemini 2.5 Pro achieves leading performance on our junior-dev evals. It was the first-ever model that solved one of our evals involving a larger refactor of a request routing backend. It felt like a more senior developer because it was able to make correct judgement calls and choose good abstractions," said Silas Alberti, Founding Team, Cognition. Gemini 2.5 Pro in action Gemini 2.5 Pro's deep understanding of code, combined with powerful reasoning, continues to make Gemini 2.5 Pro the go-to model for developers. The following cases are particularly exciting examples of how this model can be used. Video to code Gemini 2.5 Pro delivers state-of-the-art video understanding, scoring 84.8% on the VideoMME benchmark. Combining this with coding enables new flows that were previously not possible with earlier versions. For example, the Video to Learning App in Google AI Studio demonstrates how Gemini 2.5 Pro creates an interactive learning app based on a single YouTube video. With improved video understanding and complete UI, the updated Gemini 2.5 Pro model delivers a more functional experience than the previous simple example. Easier feature development Gemini 2.5 Pro is strong at front-end web development, helping developers get more done. Implementing new features means manually diving into design files and inspecting components to match style properties like colors, fonts, padding, margins, and borders then manually writing the CSS code needed to replicate those visual properties accurately. Now imagine using Gemini 2.5 Pro in an IDE and having the model generate new features, like adding a video player in the style of the other apps in the Gemini 95 starter app. Quick concepts to working apps Bringing ideas to life with both functionality and a beautiful UI is made easier with Gemini 2.5 Pro. The new dictation starter app, built using the updated model, is a great example of this in action. Some of the details such as the wavelength animations, responsive design, and subtle button hover effects stand out. By default, the model has a real taste for aesthetic web development while maintaining its steerability, helping developers quickly take a concept to a working web app. Gemini 2.5 Pro was able to design and code the microphone UI animation for the dictation starter app. Start building with Gemini 2.5 Pro Developers can build with Gemini 2.5 Pro using the Gemini API in Google AI Studio, and enterprise customers can use Vertex AI. For developers already using Gemini 2.5 Pro, this new version not only improves coding performance but also addresses key developer feedback including reducing errors in function calling and improving function calling trigger rates. The previous iteration (03-25) now points to the most recent version (05-06), so no action is required to use the improved model, and it continues to be available at the same price. The model card has also been updated with the new version of 2.5 Pro. Become a subscriber of App Developer Magazine for just $5.99 a month and take advantage of all these perks. MEMBERS GET ACCESS TO- Exclusive content from leaders in the industry - Q&A articles from industry leaders - Tips and tricks from the most successful developers weekly - Monthly issues, including all 90+ back-issues since 2012 - Event discounts and early-bird signups - Gain insight from top achievers in the app store - Learn what tools to use, what SDK's to use, and more Subscribe here Gemini Pro preview - Even better coding performance, Gemini 2.5 Pro, Agentic Workflows, Front End Development, AI Coding Tools share

Google buffs Gemini Pro just before its I/O show

Google has launched Gemini 2.5 Pro Preview (I/O edition), an updated version of its flagship Gemini 2.5 Pro AI model, ahead of its annual I/O developer conference. The new model is available via the Gemini API, Google's Vertex AI and AI Studio platforms, as well as in the company's Gemini chatbot app for web and mobile devices. The Gemini 2.5 Pro Preview (I/O edition) boasts "significantly" improved capabilities for coding and building interactive web apps. It leads the WebDev Arena Leaderboard, a benchmark that measures a model's ability to create functional and aesthetically pleasing web apps. The model also achieves state-of-the-art performance in video understanding, scoring 84.8% on the VideoMME benchmark. Video: Demis Hassabis, Co-Founder & CEO of Google DeepMind Google claims that Gemini 2.5 Pro Preview (I/O edition) is particularly adept at tasks such as code transformation and code editing. For developers already using Gemini 2.5 Pro, the new version promises to improve coding performance, reduce errors in function calling, and improve function calling trigger rates. The launch comes as Google competes fiercely with rivals like OpenAI and xAI in the AI race. The company's I/O conference is expected to feature a host of new models, AI-powered tools, and platforms. Gemini 2.5 Pro Preview (I/O edition) is priced the same as the Gemini 2.5 Pro model it replaces. Google's blog post highlights the model's capabilities, stating that it has "a real taste for aesthetic web development while maintaining its steerability." The company is positioning Gemini 2.5 Pro Preview (I/O edition) as a top contender in the AI landscape, with its improved coding capabilities and performance.

Google Is Updating Gemini 2.5 Pro Model With Vibe Coding Capabilities

It is also available in the Gemini API via Google AI Studio and Vertex AI Google updated the existing Gemini 2.5 Pro artificial intelligence (AI) model with enhanced coding capabilities on Tuesday. The Mountain View-based tech giant is calling the newer version of the model Gemini 2.5 Pro Preview (I/O edition). Interestingly, the company said that it had planned to release this version at its upcoming Google I/O 2025 event, scheduled to be held between May 20-21. However, due to positive feedback for the model, the tech giant decided to release it earlier. In a blog post, the tech giant detailed the updated large language model. Google claims that with improved code understanding, the Gemini 2.5 Pro (I/O edition) can even build interactive web apps from scratch. Apart from user interface development, the model can also perform other complex coding tasks such as code transformation, code editing, and developing complex agentic workflows. The updated model is now also ranking in the top position on the WebDev Arena leaderboard. The crowd-sourced leaderboard was developed by LMArena and measures human preference for a model's ability to complete web development tasks. The model previously ranked in second position, but has now taken the crown from Claude 3.7 Sonnet. Gemini 2.5 Pro (I/O edition) also comes with a new video-to-code feature. The model's video understanding capability helped it score 84.8 percent on the VideoMME benchmark. Now, with improved coding performance, it can create an interactive web app based on a single YouTube video. This functionality is currently only available in Google AI Studio. Additionally, the model is now also better at front-end web development. Google claims the AI model can be connected with an integrated development environment (IDE) and inspect components to match style properties such as colours, fonts, padding, margins, and more. Developers can access and begin vibe coding with the updated Gemini 2.5 Pro model in the Gemini application programming interface (API) via Google AI Studio and Vertex AI. Regular users can find the new model in the Gemini app and web client. It also powers features such as Canvas.

Google launches Gemini 2.5 Pro IO edition, optimized for coding and software development

Scored 63.8% on SWE-Bench Verified, highlighting its real-world potential in autonomous software engineering with a custom agent-based setup.Google is gearing up for its annual showcase -- Google I/O 2025, scheduled for May 20. While the event is expected to unveil major software updates like Android 16, Android XR, and new Gemini AI features, Google decided to drop a little surprise ahead of time. DeepMind CEO Demis Hassabis couldn't hold back and revealed upgrades to what he calls their "most intelligent model" yet. In his words, these are "pre-I/O goodies" -- with the highlight being that Gemini 2.5 Pro (Preview) now comes with massively improved coding capabilities. Just ahead of its big stage debut at Google I/O 2025, Google has quietly released a preview of its enhanced Gemini 2.5 Pro, now branded the I/O Edition. Internally labelled gemini-2.5-pro-preview-05-06, this version takes a big step forward, especially when it comes to coding. Google says the model now delivers significant improvements in code transformation, code editing, and even in developing complex agentic workflows -- making it far more capable for software developers and engineers. This isn't a minor bump -- it builds on the original Gemini 2.5 Pro, which launched in March, and pushes the architecture to handle more demanding and intricate programming tasks. Google calls every model in the 2.5 line a "thinking model," designed to reason through its responses, and that design is clearly paying off here. In terms of performance, Gemini 2.5 Pro (I/O Edition) now leads the LMArena leaderboard (which is based on human preference), and tops academic benchmarks like AIME 2025 for math and GPQA diamond for science -- and it does all this without relying on test-time tricks like majority voting. Where it really shines is programming. Google claims it's a "big leap over 2.0" with better support for building rich web apps, agent-driven tools, and performing advanced code edits. In the SWE-Bench Verified test for autonomous software engineering, it scored 63.8% using a custom agent-based setup -- showing it's more than ready for real-world development use cases.

Google Improves Coding Capabilities in Gemini 2.5 Pro Preview | PYMNTS.com

Google is providing early access to an updated version of its Gemini 2.5 Pro multimodal artificial intelligence (AI) model, the company said in a Tuesday (May 6) bog post. Called Gemini 2.5 Pro Preview, the model has "significantly" improved capabilities for coding, especially for interactive web applications, according to the company. Google said it released the preview model early due to "overwhelming enthusiasm" for the model "so people can start building." The model was supposed to be unveiled at Google's I/O developer conference later this month. The updates include enhancing code transformation, code editing and developing of complex agentic workflows. Google said the updates enabled the model to top the chart in the WebDev Arena Leaderboard, which ranks large language models based on their performance in web development. This ranking measures how well a model can build "aesthetically pleasing and functional" web applications. In the arena, AI models compete with each other in front-end UI design and coding contests to earn ELO points. Google said Gemini 2.5 Pro Preview has surpassed the prior version by 147 ELO points. Gemini 2.5 Pro is in third place, after Claude 3.7 Sonnet. OpenAI's GPT-4.1 came in fourth and Claude 3.5 Sonnet is fifth. Gemini 2.5 Pro Preview also scored well in video understanding, at 84.8% on the VideoMME benchmark, which assesses the capabilities of multimodal models in how well they analyze videos. Gemini is Google's flagship family of AI models and the first major natively multimodal model, meaning it was built to be able to handle text, images, code and video from the ground up. It has a 1 million token context window, which Google plans to increase to 2 million. This is the window where users enter prompts. One million tokens translates to about 750,000 words or about 10 novels. Meta's Llama 4 recently surpassed Gemini with a 10 million context window. Google unveiled Gemini 2.5 in March as its most powerful AI model yet. It is billed as a "thinking" model that is capable of reasoning instead of just classifying and predicting data. The Pro Experimental version came out first and debuted at the top of the LMArena leaderboard, where users vote on which AI chatbot generates the best responses to prompts. Gemini 2.5 Pro Preview is now at the top of LM Arena.

Google unveils early access to Gemini 2.5 Pro Preview for developers

Google on Tuesday introduced early access to Gemini 2.5 Pro Preview (I/O edition), an updated version of its Gemini 2.5 Pro model designed to improve coding tasks, particularly for creating interactive web applications. Tulsee Doshi, Senior Director of Product Management, said the update was initially planned for release at Google I/O in a few weeks. However, due to strong interest in the model, Google decided to make it available sooner to allow developers to begin building. Doshi noted that the update builds on positive feedback about Gemini 2.5 Pro's coding and multimodal reasoning abilities. The updated model leads the WebDev Arena Leaderboard, outperforming its predecessor by 147 Elo points. This leaderboard evaluates a model's ability to create visually appealing and functional web applications based on human preferences. The model also maintains strong performance in native multimodality and long-context processing, achieving an 84.8% score on the VideoMME benchmark for video understanding. Developers can use the updated Gemini 2.5 Pro via the Gemini API on Google AI Studio and Vertex AI, while the Gemini app incorporates the model to support Canvas, allowing users to create code and build interactive web applications with a single prompt.