Google has introduced Gemini 3.1 Pro, a significant upgrade to its artificial intelligence platform that signals a shift toward more autonomous and capable AI systems.
Released in preview on Thursday, the new model more than doubles the reasoning performance of its predecessor, positioning Google to accelerate the development of AI tools designed to handle complex, real-world problem solving rather than simple text generation.
The latest model expands on the architecture of the Gemini 3 series launched in November, focusing on what Google describes as “agentic” workflows. These systems are designed to independently manage multi-step tasks across domains such as software engineering, financial analysis, and academic research, reflecting a broader industry push toward AI that can plan, reason, and execute tasks with minimal human input.
According to Google’s internal benchmarks, Gemini 3.1 Pro achieved a score of 77.1 percent on ARC-AGI-2, a demanding evaluation that measures how effectively an AI model can solve unfamiliar logic problems it has not encountered during training. The company says this result represents more than double the performance of the earlier 3.0 Pro model and significantly narrows the gap with the recently introduced Deep Think specialised mode.
Among the model’s notable new capabilities is its ability to generate fully functional, website-ready animated Scalable Vector Graphics directly from a text prompt, highlighting improvements in multimodal creation and practical design applications.
The launch underscores Google’s continued investment in advancing core AI intelligence, as competition intensifies among technology companies racing to build more powerful and autonomous artificial intelligence systems.
News Source: Emirates News Agency
