On Thursday, Google released the latest model of Gemini Professional, its highly effective LLM. The mannequin, 3.1, is at present accessible as a preview and will likely be typically launched quickly, the corporate stated.
Google’s new mannequin could also be one of the vital highly effective LLMs but. Onlookers have famous that Gemini 3.1 Professional seems to be a giant step up from its predecessor, Gemini 3—which, upon its launch in November, was already thought-about a extremely succesful AI software.
On Thursday, Google additionally shared statistics from unbiased benchmarks—reminiscent of one referred to as Humanity’s Final Examination—that confirmed it performing considerably higher than its earlier model.
Gemini 3.1 Professional was additionally praised by Brendan Foody, the CEO of AI startup Mercor, whose benchmarking system, APEX, is designed to measure how properly new AI fashions carry out actual skilled duties. “Gemini 3.1 Professional is now on the prime of the APEX-Brokers leaderboard,” Foody stated in a social media post, including that the mannequin’s spectacular outcomes present “how shortly brokers are bettering at actual data work.”
The discharge comes because the AI model wars are heating up, and tech firms proceed to launch more and more highly effective LLMs designed for agentic work and multi-step reasoning. Different main names—together with OpenAI and Anthropic—have just lately launched new fashions as properly.
Techcrunch occasion
Boston, MA
|
June 9, 2026
