Key takeaways Google released Gemini 3, its most advanced AI model, now available on Vertex AI and Gemini Enterprise for business teams and developers. The model achieved a record-breaking 1501 Elo sc...
Google released Gemini 3 on Tuesday, making its most capable AI model available to enterprise customers through Vertex AI and Gemini Enterprise.
The launch comes seven months after Gemini 2.5 and positions the company in direct competition with OpenAI's GPT-5 and Anthropic's Claude Sonnet 4.5.
The model features a 1 million token context window and introduces what Google describes as state-of-the-art reasoning capabilities combined with multimodal understanding across text, images, video, and audio.
"With Gemini 3, we're seeing this massive jump in reasoning," said Tulsee Doshi, Google's head of product for the Gemini model, in an interview with TechCrunch. "It's responding with a level of depth and nuance that we haven't seen before."
Google also announced Antigravity, a new agentic development platform designed for coding tasks.
The IDE combines a prompt interface with terminal and browser windows to help developers build applications using AI assistance.
"The agent can work with your editor, across your terminal, across your browser to make sure that it helps you build that application in the best way possible," said Koray Kavukcuoglu, Chief Technology Officer at Google DeepMind.
Gemini 3 Pro achieved the top position on the LMArena Leaderboard with a score of 1501 Elo.
On the Humanity's Last Exam benchmark, which measures general reasoning and expertise, the model scored 37.4% without using external tools—exceeding GPT-5 Pro's previous record of 31.64%.
The model also scored 91.9% on GPQA Diamond, a benchmark measuring PhD-level reasoning, and set a new standard of 23.4% on MathArena Apex for mathematical problem-solving.
Google emphasized the model's reduced sycophancy compared to previous versions.
According to a statement from Demis Hassabis, CEO of Google DeepMind, AI responses powered by Gemini 3 will be "trading cliché and flattery for genuine insight—telling you what you need to hear, not what you want to hear."
Josh Woodward, Vice President of Google Labs and Gemini, told reporters that Gemini 3 is the company's "best vibe coding model ever," referring to the growing market of tools that allow developers to generate code through natural language prompts.
Several major technology companies have begun integrating Gemini 3 into their platforms.
Third-party coding tools including Cursor, GitHub Copilot, JetBrains, Replit, and Figma Make are offering access to the model.
"In our early testing in VS Code, Gemini 3 Pro demonstrated 35% higher accuracy in resolving software engineering challenges than Gemini 2.5 Pro," said Joe Binder, Vice President of Product at GitHub. "That's the kind of potential that translates to developers solving real-world problems with more speed and effectiveness."
Mikhail Parakhin, Chief Technology Officer at Shopify, described the advancement in practical terms: "Gemini 3 is a major leap forward for agentic AI. It follows complex instructions with minimal prompt tuning and reliably calls tools, which are critical capabilities to build truly helpful agents. This advancement accelerates Shopify's ability to build agentic AI tools that solve complex commerce challenges for our merchants."
Thomson Reuters is evaluating the model for legal applications. "Our early evaluations indicate that Gemini 3 is delivering state-of-the-art reasoning with depth and nuance.
We have observed measurable and significant progress in both legal reasoning and complex contract understanding," said Joel Hron, Chief Technology Officer at Thomson Reuters.
Bob Bradley, Vice President of Data Science and AI Engineering at Geotab, reported quantifiable improvements in their testing: "We immediately achieved a 10% boost in the relevancy of responses for a complex code-generation task used for data retrieval and noted a further 30% reduction in tool-calling mistakes."
Gemini 3 Pro is available now in preview on Vertex AI for developers and through Gemini Enterprise for business teams.
The model is also accessible through Gemini CLI, AI Studio, and the new Antigravity platform on Mac, Windows, and Linux.
A more advanced version called Gemini 3 Deep Think will be released to Google AI Ultra subscribers in the coming weeks after additional safety testing.
According to Google, the model achieves 41.0% on Humanity's Last Exam without tools and 93.8% on GPQA Diamond.
Key takeaways OpenAI has signed a multi-year, $10 billion agreement with AI chipmaker Cerebras Systems to secure computing infrastructure. The deal will deliver 750 megawatts of computing power throug...
KEY TAKEAWAYS: xAI implemented restrictions preventing Grok from editing images of real people in revealing clothing after global backlash California Attorney General Rob Bonta launched investigation...
Key takeaways TSMC reported fourth-quarter net income of NT$505.74 billion, beating analyst expectations and setting a new record. Fourth-quarter revenue reached NT$1.046 trillion, up 20.5% year over...
