Google unveiled its latest iteration of the formidable Gemini Pro large language model, version 3.1, on Thursday. The company stated that this model is presently accessible as a preview, with a broader general release anticipated shortly.
This new model from Google is poised to be among the most potent large language models developed to date. Industry observers have remarked that Gemini 3.1 Pro represents a substantial advancement over its precursor, Gemini 3, which itself was recognized as an exceptionally capable AI tool when it debuted in November.
Furthermore, on Thursday, Google presented independent benchmark data, including results from a test known as "Humanity’s Last Exam," which demonstrated Gemini 3.1 Pro's markedly superior performance compared to its earlier iteration.
The capabilities of Gemini 3.1 Pro also garnered commendation from Brendan Foody, CEO of the AI startup Mercor. Foody's company developed the APEX benchmarking system, which is specifically engineered to assess the efficacy of new AI models in executing genuine professional tasks. In a social media post, Foody declared, “Gemini 3.1 Pro is now at the top of the APEX-Agents leaderboard,” further noting that the model's remarkable performance underscores “how quickly agents are improving at real knowledge work.”
This launch occurs amidst an intensifying competitive landscape in AI model development, as technology giants persist in introducing progressively sophisticated large language models engineered for autonomous "agentic" operations and complex multi-step reasoning. Prominent industry players, such as OpenAI and Anthropic, have also recently unveiled their own new model advancements.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.