Google has officially debuted Gemini 3, its latest and most powerful artificial intelligence model. The new model, described by executives as a major step toward Artificial General Intelligence (AGI), boasts state-of-the-art reasoning, multimodal understanding, and advanced coding capabilities, and is being immediately integrated across Google's entire product ecosystem.
In a series of announcements, Google and Alphabet CEO Sundar Pichai framed the launch as the culmination of years of intense effort. "Nearly two years ago we kicked off the Gemini era, one of our biggest scientific and product endeavors ever undertaken as a company," Pichai noted.
Gemini 3 is positioned as Google's most intelligent model yet, designed to grasp complex concepts and user intent with depth. This focus on contextual awareness is a core selling point. Pichai emphasized that the model is "much better at figuring out the context and intent behind your request, so you get what you need with less prompting. It’s amazing to think that in just two years, AI has evolved from simply reading text and images to reading the room."
Benchmark Breakthroughs and Enhanced Capabilities
The model, specifically the Gemini 3 Pro version, has delivered remarkable results across major industry benchmarks, significantly outperforming its predecessor, Gemini 2.5 Pro. Key upgrades include:
- Reasoning: Gemini 3 Pro achieved a 37.5 per cent score (without tools) on Humanity’s Last Exam and 91.9 per cent on GPQA Diamond, demonstrating "PhD-level reasoning."
- Multimodality: It set new standards for understanding across different formats, scoring 81 per cent on MMMU-Pro and 87.6 per cent on Video-MMMU, solidifying its position as "the best model in the world for multimodal understanding," according to DeepMind CEO Demis Hassabis and Chief AI Architect Koray Kavukcuoglu.
- Coding & Agentic Tools: The model showed exceptional performance in autonomous capabilities, topping the WebDev Arena leaderboard with 1487 Elo and achieving 76.2 per cent on SWE-bench Verified.
For advanced users, Google is also introducing "Gemini 3 Deep Think" mode, an enhanced reasoning feature that pushed scores even higher during internal testing.
Integration into Search and New Developer Tools
Google is immediately rolling out Gemini 3 into its core products. Crucially, this generation is the first to be available in Google Search on its launch day.
"And starting today, we’re shipping Gemini at the scale of Google. That includes Gemini 3 in AI Mode in Search with more complex reasoning and new dynamic experiences. This is the first time we are shipping Gemini in Search on day one," Pichai said. The integration in Search will unlock dynamic, generative user interfaces, interactive tools, and simulations that are generated on the fly to help users "make better sense of information on the web."
For developers, Google introduced Google Antigravity, a new agentic development platform that leverages Gemini 3's advanced coding and tool-use capabilities, allowing agents to autonomously plan and execute complex software tasks.
The Competitive Landscape
The Gemini 3 launch intensifies the rivalry with OpenAI, which recently issued two updates to its GPT-5 model. OpenAI described these updates as making the model “warmer, more intelligent, and better at following your instructions,” and noted that one variant is "faster on simple tasks, more persistent on complex ones.”
Google's user base for its existing AI products suggests high stakes in this competition. The company reported that AI Overviews now have 2 billion users every month, and the Gemini app surpasses 650 million users per month. This scale provides Google with a powerful distribution advantage as it seeks to maintain its dominance in search and cloud services against the backdrop of rapidly evolving AI capabilities. |
