Google updated its Gemini 2.5 Pro preview model with better coding and reasoning skills. The refreshed version dropped Thursday and is already live on AI Studio, Vertex AI, and the Gemini app. Full release is expected in a couple of weeks.
Google says Gemini 2.5 Pro “continues to excel at coding” and leads “difficult coding benchmarks.” The model also scored top marks on tests covering math, science, knowledge, and reasoning.
The new build improves response style and structure, addressing feedback from last month’s 2.5 Pro release. Now it’s “more creative with better-formatted responses,” according to Google. 
Google wrote in its blog:
“[Gemini 2.5 Pro] continues to excel at coding, leading on difficult coding benchmarks,” Google
wrote in a blog post. “It also shows top-tier performance [on] highly challenging benchmarks that evaluate a model’s math, science, knowledge, and reasoning capabilities.”
