Google Improves Coding Capabilities in Gemini 2.5 Pro Preview
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.
Read more: Google Debuts Touted Gemini 2.5 in the ‘Winner-Take-All’ AI Model Race
The post Google Improves Coding Capabilities in Gemini 2.5 Pro Preview appeared first on PYMNTS.com.