On Thursday, Google introduced the latest iteration of its advanced language model, Gemini Pro 3.1, now available in a preview version with a full release on the horizon.
This new model is being hailed as one of the most formidable language models to date. Observers note that Gemini 3.1 Pro marks a significant enhancement over its predecessor, Gemini 3, which was already recognized as a highly effective AI tool upon its launch last November.
Google also released independent benchmark results, including a notable assessment known as Humanity's Last Exam, showcasing Gemini 3.1 Pro's impressive performance compared to earlier versions.
Brendan Foody, CEO of the AI startup Mercor, praised the model, stating that it currently leads the APEX-Agents leaderboard, a benchmarking system that evaluates AI models on their ability to perform real-world professional tasks. He emphasized that the results reflect the rapid advancements in agents' capabilities for knowledge work.
The launch of Gemini 3.1 Pro occurs amidst a competitive landscape in AI development, as tech giants continue to unveil increasingly sophisticated language models aimed at enhancing agentic tasks and complex reasoning. Companies like OpenAI and Anthropic have also recently introduced their own advanced models.