Latest version brings improvements in creativity, formatting, and advanced benchmark performance
June 6, 2025 — TechXNow.com
Google has unveiled a new update to its Gemini 2.5 Pro model, which the company says delivers significant enhancements in coding performance and overall response quality. Announced on Thursday, this “updated preview” version is now live on Google’s AI platforms — AI Studio, Vertex AI, and the Gemini app — with broader general availability expected in the coming weeks.
Refining a Flagship Model
The Gemini 2.5 Pro model, part of Google’s broader Gemini family of generative AI systems, has already gained recognition for its capabilities across programming, reasoning, and complex problem-solving. According to Google, the latest update is not a full version jump but a critical refinement based on user feedback from the earlier Gemini 2.5 Pro release.
In an official blog post, Google said the model “continues to excel at coding” and now ranks among the top performers on difficult coding benchmarks. These benchmarks test a model’s ability to handle advanced programming tasks with precision and logical structure.

Boosted Coding and Reasoning Abilities
Beyond its improvements in code generation, the updated Gemini 2.5 Pro also delivers strong results in areas such as mathematics, science, general knowledge, and logical reasoning. Google emphasized that the model ranks highly on “highly challenging benchmarks” that evaluate a wide spectrum of cognitive abilities — making it more versatile for developers, researchers, and enterprise users alike.
These enhancements further position Gemini 2.5 Pro as a competitive alternative to other leading AI models in the market, especially for technical tasks that require both accuracy and adaptability.
ALSO SEE : DeepSeek May Have Used Google Gemini Data for AI Model
Focus on Style, Structure, and Creativity
One of the standout improvements in this update is the refinement of the model’s output style. Google reports that Gemini 2.5 Pro now produces better-formatted responses with a greater degree of creativity and clarity. This means users can expect more readable code snippets, cleaner documentation, and responses that follow a consistent logical structure.
The company says it actively incorporated community and developer feedback to shape this update, improving how the model communicates results, writes structured content, and adapts its tone for different use cases.

Now Available for Developers
As of today, the updated Gemini 2.5 Pro model is accessible via Google’s AI Studio and Vertex AI platforms, as well as through the Gemini app. This gives developers and AI enthusiasts early access to test and integrate the improved model into their workflows.
Full-scale general availability is expected in the coming weeks, according to Google.
Stay tuned to TechXNow for further updates on Gemini developments and the broader generative AI landscape.
Sources : ( Techcrunch )