Editorial illustration for Gemini 3.1 Pro released as upgraded Gemini 3 Deep Think for complex tasks
Gemini 3 Deep Think: AI Breakthrough in Science & Code
Gemini 3.1 Pro released as upgraded Gemini 3 Deep Think for complex tasks
The useful AI doesn't just draft your email. It solves the problem you couldn't crack. Think calculator versus physicist.
Google aimed its Gemini 3 Deep Think system at brutal scientific challenges just last week. Now, that same core intelligence gets a wider release and a new name: Gemini 3.1 Pro. The logic is straightforward.
Google is installing it as the standard brain across its entire toolkit, from API right down to the consumer app. A model capable of aiding drug design, the thinking goes, should also be the one debugging your complex code.
Last week, we released a major update to Gemini 3 Deep Think to solve modern challenges across science, research and engineering. Today, we’re releasing the upgraded core intelligence that makes those breakthroughs possible: Gemini 3.1 Pro. We are shipping 3.1 Pro across our consumer and developer products to bring this progress in intelligence to your everyday applications.
Forget shiny new features. This is about raising the floor. The reasoning engine from last week's specialized demo is now the default.
It's live in Vertex AI for enterprises. Developers can call it from the API. It's inside the Gemini app on your phone.
Google's pitch is that simple. The move signals a clear bet: to compete, you don't need a zoo of niche models. You need one profoundly smarter generalist.
Can the thinking that tackles nuclear fusion also untangle a broken spreadsheet? We'll see.
Related Reading
Browse all Google news →
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
Related Topic NVIDIA and Google Cloud let developers scale AI from prototype to production
Market Trends
Related Topic Google's FACTS benchmark shows 70% factuality ceiling across four tests
Research & Benchmarks
From Archives Google embeds Lyria, expanding AI music beyond niche platforms Suno, Udio
From Archives Google adds Lyria 3 text‑to‑music to Gemini, enabling song creation via prompts
From Archives Google launches AI Professional Certificate to boost fluency for workers
Research & Benchmarks
From Archives Gemini app expands tools, now generates music alongside images and video
Common Questions Answered
What makes Gemini 3 Pro different from previous Gemini models?
Gemini 3 Pro is positioned as Google DeepMind's most intelligent system, significantly outperforming previous versions across major benchmarks. Unlike older models, it is built to work across text, images, audio, video, and code, with the ability to orchestrate tools and apps to complete complex tasks.
How does Gemini 3 Pro perform on advanced reasoning benchmarks?
Google reports that Gemini 3 Pro significantly outperforms Gemini 2.5 Pro on every major benchmark, particularly in reasoning and factual accuracy. On the Humanity's Last Exam (HLE), the Deep Think mode reaches around 41%, and on the GPQA Diamond graduate-level question answering benchmark, it achieves approximately 93-94% accuracy.
What unique capabilities does Gemini 3 Pro offer in terms of context processing?
Gemini comes standard with a remarkable 1-million-token context window, which is a significant advancement over previous language models that were limited to 8,000-128,000 tokens. This long context window enables processing of extensive amounts of information, such as over 200 podcast transcripts, 8 novels, or 50,000 lines of code, with near-perfect retrieval (>99%).
Further Reading
- A new era of intelligence with Gemini 3 - The Keyword — Google Blog
- Gemini 3 Deep Think: Advancing science, research and engineering — Google Blog
- Gemini Deep Think: Redefining the Future of Scientific Research — DeepMind Google Blog
- Release notes | Gemini API - Google AI for Developers — Google AI for Developers
