Gemini 3.1 Pro vs Gemini 3 Pro: Google’s new AI is slower on purpose (www.techradar.com)

Google has unveiled Gemini 3.1 Pro, an upgraded version of its AI model that focuses on enhanced reasoning capabilities, albeit at the cost of speed in certain modes. Unlike its predecessor, Gemini 3 Pro, which often rushed to plausible answers, the new model introduces a "deep think mode" that has demonstrated significant improvements in complex reasoning tasks, as evidenced by higher scores on benchmarks like ARC-AGI-2. Notably, Gemini 3.1 Pro can adeptly manage Scalable Vector Graphics (SVGs), enabling it to write and animate code with precision. The significance of Gemini 3.1 Pro lies in its deepened understanding of intricate prompts, thereby elevating its potential for applications that require advanced logical structures. For example, in a test scenario involving abstract reasoning about liquid behavior in a fictional gravity scenario, the new model excelled where the previous one faltered, generating accurate descriptions and solutions. Similarly, when tasked with creating an animated SVG solar system and planning logistics for a fictional supervillain operation, Gemini 3.1 showcased superior narrative coherence and intricate planning while embracing a creative persona. This evolution reinforces the necessity of adopting Gemini 3.1 Pro for tasks demanding complex, multifaceted reasoning, distinguishing itself from merely adequate models like Gemini 3 Pro.