Google just upgraded Gemini again to 3.1 Pro (www.techradar.com)

Google has launched the enhanced Gemini 3.1 Pro, significantly improving AI reasoning performance, more than doubling the problem-solving capabilities of its predecessor, Gemini 3 Pro. The new model achieved a score of 77.1% on the ARC-AGI-2 benchmark, designed for evaluating logic-solving abilities. This upgrade is now accessible across various subscription tiers, including Free, Plus, Pro, and Ultra, and is also integrated into Google’s NotebookLM research tool. One of the applications demonstrated is a city planner tool that handles complex terrain, infrastructure mapping, and traffic simulation, showcasing Gemini 3.1 Pro's enhanced multimodal reasoning. While the technical advancements are notable, user feedback reveals a concern about a perceived reduction in the model's emotional depth and empathy. Some users feel that the upgrade, while impressive in logic and problem-solving, has made the AI's interactions feel less human and nuanced, echoing early criticisms faced by similar upgrades in other AI systems. Google acknowledges the mixed reactions and reassures users that 3.1 Pro remains a preview version, with ongoing developments aimed at improving emotional intelligence and agentic workflows in future updates.