Google has unveiled its latest advancement in artificial intelligence with the launch of Gemini 3.1 Pro, an upgraded version of its large language model (LLM). Currently available in preview mode, the full release is expected shortly, signaling a significant enhancement over its predecessor, Gemini 3, which received acclaim when introduced last November.
Gemini 3.1 Pro has already garnered attention for achieving exceptional scores in independent benchmarks, including the Humanity’s Last Exam. These results indicate a substantial performance improvement over its earlier variant. Brendan Foody, CEO of AI startup Mercor, emphasized the model’s capabilities by declaring it the leader on the APEX-Agents leaderboard, a platform designed to assess AI models in real-world professional tasks. Foody noted that the remarkable results illustrate the rapid advancements in AI agents’ proficiency in knowledge work.
This release intensifies competition in the AI model market, as tech giants like OpenAI and Anthropic also roll out their latest models targeted at complex reasoning and agentic tasks.
Key Highlights:
– Google launches Gemini 3.1 Pro, an advanced LLM.
– Achieves record benchmark performance, surpassing Gemini 3.
– Brendan Foody recognizes it as a top performer in real-world task assessments.
– The race among AI firms intensifies with recent releases from major competitors.
