Google’s launch of Gemini 3 is changing the landscape of artificial intelligence. It has received very positive feedback from some of the industry’s top leaders. Salesforce CEO Marc Benioff is among the strongest supporters. This week, he referred to the new model as a significant advancement for Google, one that has convinced him to move away from ChatGPT, which he has used for a long time.
Feedback from early testers, benchmarks shared during Google’s demo, and statements from tech executives suggest that Gemini 3 may be Google’s best entry in the AI competition since OpenAI made a splash three years ago.
Benioff’s Bold Endorsement: “I’m Not Going Back to ChatGPT”
In a post on Monday, Benioff shared that after spending just a few hours with Gemini 3, he didn’t plan to return to ChatGPT. He listed three key areas where the new model seems much better:
- Speed
- Reasoning abilities
- Creative output, especially in images and video
“It’s an extraordinary step forward,” Benioff wrote, reflecting the excitement shared privately by several enterprise leaders who have tried the system.
The Salesforce CEO also pointed to a Wall Street Journal analysis indicating that Google may have surpassed its closest competitors, OpenAI and Anthropic, with this release. This endorsement is significant because Salesforce integrates multiple AI models into its Einstein platform, and Benioff has been an active user and critic of emerging technologies.
A Model That Impresses Beyond Benchmarks
Benioff’s praise fits into a larger trend of positive early feedback. During Google’s prominent demo, the company highlighted new performance metrics that showed Gemini 3 outperforming its competitors in several reasoning and multimodal tests. However, the real-world responses from early users seem to resonate even more.
Tulsee Doshi: Improvements in Underrepresented Languages
Tulsee Doshi, Google’s Senior Director of Product Management for Gemini, provided a telling example of Gemini 3’s enhanced capabilities. She asked the model to write in Gujarati, a popular Indian language that has limited representation online.
The results were “noticeably better” than earlier versions, she said—indicating Gemini 3 is addressing long-standing gaps in its ability to handle multiple languages. For global enterprises, especially those in India, Africa, or Southeast Asia, this improvement could expand usage significantly.
Early Access Evaluations Show “Striking Gains”
Apart from Google’s own announcements, independent testers report significant improvements in various technical areas.
Box CEO Aaron Levie: Performance Jump Was “So Big We Rechecked Our Tests”
Aaron Levie, CEO of the cloud content management company Box, noted that his team ran early evaluations with large and complex document sets. He said Gemini 3 produced results that were so strong that engineers questioned whether they had set up their benchmark tools correctly.
Levie described the performance gains as “surprisingly large,” marking one of the biggest jumps his team has seen across different models.
These reactions are important because Box supports enterprise document workflows for tens of thousands of companies. This makes its tests a useful indicator of real-world AI tasks, including summarization, extraction, classification, and reasoning over long contexts.
A Rare Advantage for Google in the AI Race
For Google, Gemini 3 is a long-awaited chance to regain momentum after years of lagging behind OpenAI’s ChatGPT and Anthropic’s Claude in public perception.
The last three years have been tough for Google:
- OpenAI dominated early generative AI adoption.
- ChatGPT became the global standard for consumers and professionals.
- Anthropic established a solid presence among developers and technical users.
However, the initial reactions to Gemini 3 represent a rare moment where Google seems to have taken control of the narrative. Analysts and executives believe the model offers advantages in:
- Complex reasoning and logic tasks
- Multimodal generation, particularly in images and videos
- Speed and responsiveness
- Creative writing and content generation
- Lower latency for enterprise-scale tasks
Michael Nathanson, an analyst at MoffettNathanson, stated that Google now appears to be “firmly in the lead,” at least for this development cycle, calling Gemini 3 the “clear winner” among recent model releases.
Gemini 3 to Power New Image Tools, Including Nano Banana Update
Google has also confirmed that its new model will support an upgraded version of Nano Banana, a popular image generation tool within the Gemini ecosystem.
Nano Banana has played a quiet but significant role in driving up Gemini’s usage numbers in 2024 and 2025. This is especially true among creators, marketers, and developers who experiment with quick visual prototyping. With Gemini 3 powering the next version, Google expects further increases in adoption.
ChatGPT vs. Claude vs. Gemini: A Competitive Landscape Shifts
The release of Gemini 3 comes during a highly competitive period among AI leaders. Each platform currently excels in different areas:
- ChatGPT (OpenAI): Consumer Scale and Accessibility
OpenAI leads in overall usage, reporting around 800 million weekly users, a staggering figure that keeps ChatGPT at the forefront of mainstream AI engagement. - Gemini (Google): Rapid Growth and Multimodal Strength
Google recently announced it has 650 million monthly users, a significant and rapidly growing user base driven by Android integration, workplace tools, and global search features. - Claude (Anthropic): A Developer Favorite
Anthropic’s Claude is highly regarded for coding tasks, long-form reasoning, and safety features tailored for enterprises. Many engineers prefer Claude for software development tasks.
However, analysts now say that Gemini 3 expands Google’s strengths across all major categories, equipping the company with one of the most versatile and competitive models available.
Inside Google’s Restructured AI Strategy
Gemini 3 is not just a new model release; it is the result of a major overhaul of Google’s AI organization.
Over the past year, Google leadership, led by CEO Sundar Pichai, has reorganized teams, streamlined research efforts, and worked to eliminate internal barriers that slowed development in the past.
Sergey Brin’s Return to Hands-On Involvement
One surprising development has been the increased involvement of Google co-founder Sergey Brin, who has returned to a more active operational role.
Brin’s hands-on engagement reportedly includes:
- Guiding research teams
- Assisting with model design choices
- Collaborating with product leaders on rollout strategies
Insiders say Brin’s return has positively affected engineering morale and product development speed.
What Gemini 3 Means for the Future of the AI Race
While the competitive landscape continues to change, the launch of Gemini 3 is seen as a key moment. For the first time since ChatGPT launched, a major model release seems to have shifted the sentiment in Google’s favor.
Enterprises now have stronger reasons to look beyond OpenAI, especially if Gemini 3 continues to do better in:
- Multimodal tasks
- Multilingual generation
- Speed and efficiency
- Enterprise-scale document reasoning
- Real-time content creation
With strong endorsements from leaders like Benioff and Levie, as well as analysts suggesting that Google may have regained the lead, Gemini 3 stands out as one of Google’s most significant AI releases to date.
The next few months will show whether this momentum lasts, but for now, it is clear that attention in the industry is shifting. As Benioff plainly stated, he’s “not going back to ChatGPT anytime soon.”
