New AI Models Take Center Stage at TechCrunch Event in San Francisco

Maxwell Zeff, a senior reporter at TechCrunch, is covering the developments in artificial intelligence at a key event taking place in San Francisco from October 27-29, 2025. Famous for his deep knowledge of AI, Zeff has written for such outlets as Gizmodo, Bloomberg, and MSNBC. He’s written on some of the biggest issues, including the…

Lisa Wong Avatar

By

New AI Models Take Center Stage at TechCrunch Event in San Francisco

Maxwell Zeff, a senior reporter at TechCrunch, is covering the developments in artificial intelligence at a key event taking place in San Francisco from October 27-29, 2025. Famous for his deep knowledge of AI, Zeff has written for such outlets as Gizmodo, Bloomberg, and MSNBC. He’s written on some of the biggest issues, including the rapid advancement of AI technologies. Most recently he has been covering the crisis involving Silicon Valley Bank.

The TechCrunch event has become a key launching pad for the world’s largest technology companies to announce and demonstrate their most recent innovations. Hardest working – one of the highlights features OpenAI researcher Noam Brown. He shared a previously unreleased AI model that was instrumental in achieving a gold medal at the International Math Olympiad (IMO). This multi-agent system illustrates the potential for collaborative AI to address nuanced issues. The move serves to further deepen the national discussion on AI use in competitive settings.

In an increasingly competitive environment, Elon Musk’s xAI has released Grok 4 Heavy, their own multi-agent system. This model garnered a lot of attention for its industry-leading performance on various benchmarks. It heralds a major step change in AI performance.

Google has stolen the attention, somewhat predictably, with their announcement of Gemini 2.5 Deep Think. This complex multi-agent model purports to reach state-of-the-art performance on Humanity’s Last Exam (HLE). Gemini 2.5 Deep Think was first announced at Google I/O 2025 in May. It wants to change how AI contributes to addressing issues across disciplines and sectors.

A limited number of testers will be able to use it via the Gemini API. This is a smart move on Google’s part, as it shows their willingness to sharpen their AI technologies through hands-on applications with real-world feedback.

“Deep Think can help people tackle problems that require creativity, strategic planning and making improvements step-by-step,” – Google

Zeff’s eye towards these advancements is especially timely as he’s recently made it a mission to cover the quickly changing world of AI. His knowledge and understanding of these technologies will help inform their impact on industry and society as a whole. In his leisure time, you can find him hiking, biking, or otherwise enjoying the Bay Area’s vibrant food scene. In addition to this impressive outreach, he is simply an excellent reporter.

Whether you are a nonprofit, government, or industry stakeholder in the technology sector, all eyes are on this event as it unfolds. Generative AI tools such as Gemini 2.5 Deep Think and Grok 4 Heavy are taking the world by storm. They illustrate the astonishing development of artificial intelligence and its increasing use across all disciplines.