Following the creation of a new “open” AI model with better performance on a single GPU, Google has introduced an AI model update for its products – Gemini 2.5 – that combines a “significantly improved base model with enhanced post-training” for better overall performance. The company claims that the first version of Gemini 2.5 Pro, an experimental version, outperforms OpenAI, Anthropic, xAI, and DeepSeek in general AI tests that measure comprehension, math, coding, and other abilities. The new model is available in Google AI Studio or for Gemini Advanced subscribers in the model drop-down menu in the app.
The company also touts Gemini’s built-in multimodality as an advantage, as it is able to interpret not only text but also audio, still images, video, and code, and says a 2 million token context window is “coming soon” to help it process more data. Google DeepMind CEO Demis Hassabis called Gemini 2.5 Pro “a stunning state-of-the-art model, #1 on LMArena with a whopping +39 ELO points, with significant improvements in multimodal reasoning, coding, and STEM” in a post on X.
Google claims that it has made a leap in quality because Gemini models are now “reasoning” models that process tasks step by step and make more informed decisions, which they say leads to better answers and reactions to complex prompts. The blog post reads: “…we are building these thinking capabilities directly into all of our models so that they can solve more complex problems and support even more capable, context-aware agents.”
