Google’s newest AI model, Gemini 3, has been out for barely a week, yet it has already sparked major waves across the tech industry. The model, which Google positions as its most advanced and capable yet, is quickly earning praise — and one of the most vocal admirers is Salesforce CEO Mark Benioff.

Benioff, who has openly relied on ChatGPT for years, says the leap in performance he experienced with Gemini 3 was so dramatic that he’s ready to switch tools entirely. Sharing his reaction on X, Benioff wrote that after trying Gemini 3 for just two hours, he felt there was no turning back. “I’ve used ChatGPT every day for 3 years. Just spent 2 hours on Gemini 3. I’m not going back. The leap is insane — reasoning, speed, images, video everything is sharper and faster. It feels like the world just changed, again.”

His praise comes after a hands-on session with Google’s flagship AI system, which blends advanced text, image and video processing into a single unified model. Benioff highlighted Gemini 3’s rapid responses, improved reasoning, and more polished multimodal outputs, describing the overall experience as “transformative.” His post quickly went viral, drawing millions of views and igniting intense debate among AI enthusiasts, developers and corporate users.

While many echoed his excitement, not everyone was swept up in the enthusiasm. Some users questioned whether any single AI model can sustain such a dramatic edge in an industry that evolves at a breathtaking pace. One user responded pointedly to the hype, saying, “The hype cycle is exhausting. Every model trains on the same benchmarks and improves incrementally every six months doesn’t this become a moot point?”

Despite the skepticism, early reactions from industry leaders suggest that Gemini 3 is indeed shaping up to be a major contender in the current AI landscape. Even OpenAI CEO Sam Altman, a direct competitor, acknowledged its strengths, calling it a “great model.” Former Tesla AI chief Andrej Karpathy also weighed in, saying he had a “positive early impression,” and noting that it has “very solid daily-driver potential” and belongs in the “tier-1 LLM” category.

Google CEO Sundar Pichai describes Gemini 3 as the company’s “most intelligent model” yet, built to merge and upgrade all previous Gemini capabilities. The standout feature, Google says, is its high-level multimodality — the ability to move seamlessly between text, audio, images and video. This sophisticated interaction enables the model to analyse content across formats simultaneously, leading to better reasoning and richer outputs.

A major technical advantage is its massive one-million-token context window, which lets Gemini 3 process extremely large documents, full code repositories, or hours-long transcripts in a single run. Google claims this will unlock new possibilities in enterprise environments, such as more accurate financial modelling, long-horizon planning, supply-chain optimisation and advanced coding assistance. The company also highlights improvements in tool use and agentic workflows, positioning Gemini 3 as a strong choice for complex, real-world business tasks.

As the AI race intensifies, Benioff’s bold endorsement adds to the growing intrigue surrounding Google’s latest creation — and hints that Gemini 3 may indeed be setting a new benchmark in the world of intelligent models.