Asger Risom
Google’s launch of Gemini 3 has set off one of the clearest shifts yet in the rivalry between the two dominant AI labs.
Google’s launch of Gemini 3 has set off one of the clearest shifts yet in the rivalry between the two dominant AI labs. While the industry has grown used to fast model cycles and bold claims, reactions to Gemini 3 suggest something deeper: longtime ChatGPT loyalists — including high-profile builders — say Google may have taken the lead.
A decisive break
Salesforce CEO Marc Benioff, a daily ChatGPT user for three years, stunned the industry with a blunt verdict after testing Gemini 3 for two hours. “Holy shit… I’m not going back,” he wrote on X, calling the jump in reasoning, speed, and multimodal fluency “insane.” His comments spread quickly and added fuel to a debate that has defined the AI race since late 2022: who is actually ahead?
The enthusiasm matters less because of who said it, and more because of what it implies — a shift in perceptions at the very top of the enterprise AI market. Benioff’s companies partner with multiple AI vendors, including OpenAI, meaning his reaction is not tied to product loyalty.
Growing industry momentum
Early technical voices have echoed the praise. OpenAI CEO Sam Altman congratulated Google, calling Gemini 3 “a great model,” while former Tesla AI director Andrej Karpathy described it as a “tier 1 LLM” and a strong “daily driver.” Stripe CEO Patrick Collison highlighted its ability to generate interactive visual summaries, saying he was “impressed.”
Google and DeepMind describe the system as their most capable “agentic” model, combining text, vision, audio, and code under a single framework and deeply embedding it across the Google ecosystem — from Search to Workspace tools to developer APIs.
A tightening race
Gemini 3 enters the market as OpenAI rolls out ChatGPT 4.5 Turbo and prepares GPT-5, while Anthropic pushes Claude 3.5 into the same tier of high-end reasoning systems. For the first time since OpenAI’s breakthrough with GPT-4, the competitive gap appears ambiguous — and possibly inverted.
Benioff’s shift is notable because enterprises often lock into one provider’s workflows. A public declaration that Gemini 3 now feels meaningfully better will be closely watched by CIOs deciding where to place next year’s AI budgets.
Why it matters
For Google, Gemini 3 is part of a multi-year effort to reclaim leadership in generative AI after ChatGPT’s surprise dominance. For OpenAI, the model represents the benchmark its upcoming releases must beat. And for customers, the emerging parity means more volatility, faster iteration, and a competitive landscape where the “best model” may now be a moving target from month to month.
