正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Yesterday at #GoogleIO, we introduced Gemini 2.5 Pro Deep Think 🧠, pushing the frontiers of AI reasoning. This enhanced reasoning mode is built to tackle drastically complex problems – like USAMO problems that stumped previous models. Super proud of the GDM team for this one!

Oriol Vinyals

205,176 subscribers

86,860 次观看 • 1 年前 •via X (Twitter)

科学技术新闻政治 #GoogleIO

Anya Rossi• Live Now

Private livecam show

11 条评论

Oriol Vinyals 的头像

Oriol Vinyals1 年前

Deep Think uses new techniques, allowing the model to consider multiple hypotheses before responding. It's not just about math; it also gets an impressive score on LiveCodeBench 💻, MMMU & more! Learn more:

Oriol Vinyals 的头像

Oriol Vinyals1 年前

Because we're defining the frontier with Deep Think, we're taking extra care. This includes first getting input from safety experts and trusted testers🛡️ via the Gemini API to gather crucial feedback before wider release.

Coral AI News 的头像

Coral AI News2 年前

Coral AI is the most powerful AI for documents. See the difference yourself:

Permaximum Judicium 的头像

Permaximum Judicium1 年前

Can you confirm it's not just about generating multiple CoTs all the way to the end and selecting the best one via majority/consensus voting or another instance of the model itself? What the gif above suggests it generates a lot of CoTs in parallel but the model (likely another instance) eliminates/kills unpromising ones at various depths of the CoT until the most promising one is left in order to efficiently increase the accuracy while trying to save compute by not waiting for all CoTs to finish.

infrecursion 的头像

infrecursion1 年前

Isn't deep think just your version of o1-pro? How expensive is this model?

NewAIWorld 的头像

NewAIWorld1 年前

@demishassabis We're reaching new territory with AI. Already in 2025 AI intelligence will surpass most humans. But what comes next? I think we can't even imagine right now anymore.

LuisHer 的头像

LuisHer1 年前

They can't abandon those of us in the Advanced plan, the March version was very good and suddenly this May version made everything worse. And with this mega plan we can't do anything. Gemini is not only used by programmers, we use it to write for many things. don't forget them!

Reji Modiyil 的头像

Reji Modiyil1 年前

@GoogleDeepMind @GoogleDeepMind, impressive advancements. the future of ai reasoning seems incredibly bright.

Angel Suero 的头像

Angel Suero1 年前

@GoogleDeepMind @grok is this on the pro model? Or ultra?

Tom Bennet 的头像

Tom Bennet1 年前

USAMO problems? Now *that's* a challenge. Bet it still can’t negotiate a decent deal on a used car. 🤣

Peter Yang 的头像

Peter Yang1 年前

I'd love to try Deep Think, how to get access? Doing a post comparing all the reasoning models.

相关视频

We’re pushing the boundaries of intelligence even further with Gemini 3 Deep Think. 🧠 This mode meaningfully improves reasoning capabilities by exploring many hypotheses simultaneously to solve problems. Here’s how it coded a simulated dominoes game from a single prompt ⬇️

We’re pushing the boundaries of intelligence even further with Gemini 3 Deep Think. 🧠 This mode meaningfully improves reasoning capabilities by exploring many hypotheses simultaneously to solve problems. Here’s how it coded a simulated dominoes game from a single prompt ⬇️

Google DeepMind

248,035 次观看 • 8 个月前

We’ve upgraded our specialized reasoning mode Gemini 3 Deep Think to help solve modern science, research, and engineering challenges – pushing the frontier of intelligence. 🧠 Watch how the Wang Lab at Duke University is using it to design new semiconductor materials. 🧵

We’ve upgraded our specialized reasoning mode Gemini 3 Deep Think to help solve modern science, research, and engineering challenges – pushing the frontier of intelligence. 🧠 Watch how the Wang Lab at Duke University is using it to design new semiconductor materials. 🧵

Google DeepMind

3,196,473 次观看 • 6 个月前

Joscha Bach says current AI doesn't need to be conscious to execute tasks, but it must simulate reality to understand it LLMs don't just predict words To solve complex problems, they build deep internal models of spatial reasoning and the human theory of mind "that is why AI can talk about reasoning and mental states"

Joscha Bach says current AI doesn't need to be conscious to execute tasks, but it must simulate reality to understand it LLMs don't just predict words To solve complex problems, they build deep internal models of spatial reasoning and the human theory of mind "that is why AI can talk about reasoning and mental states"

Haider.

17,743 次观看 • 5 个月前

Chain-of-thought reasoning is a powerful tool to enable language models to work through complex problems. Can we use this with robots? With embodied chain-of-thought, vision-language-action (VLA) models can think through perception and planning! A 🧵👇

Chain-of-thought reasoning is a powerful tool to enable language models to work through complex problems. Can we use this with robots? With embodied chain-of-thought, vision-language-action (VLA) models can think through perception and planning! A 🧵👇

Sergey Levine

30,479 次观看 • 2 年前

Today, we’re rolling out an updated Deep Think mode available in the Gemini app for Google AI Ultra subscribers. Here’s what you need to know: — Gemini 3 Deep Think mode pushes the boundaries of intelligence even further, delivering meaningful improvement in reasoning capabilities — It works by using advanced parallel thinking to explore multiple hypotheses simultaneously — This builds on the Gemini 2.5 Deep Think variants that recently achieved a gold-medal standard at this year's International Mathematical Olympiad and at the International Collegiate Programming Contest World Finals Ultra subscribers can try Deep Think today by selecting “Deep Think” in the prompt bar and Gemini 3 Pro in the model selector.

Today, we’re rolling out an updated Deep Think mode available in the Gemini app for Google AI Ultra subscribers. Here’s what you need to know: — Gemini 3 Deep Think mode pushes the boundaries of intelligence even further, delivering meaningful improvement in reasoning capabilities — It works by using advanced parallel thinking to explore multiple hypotheses simultaneously — This builds on the Gemini 2.5 Deep Think variants that recently achieved a gold-medal standard at this year's International Mathematical Olympiad and at the International Collegiate Programming Contest World Finals Ultra subscribers can try Deep Think today by selecting “Deep Think” in the prompt bar and Gemini 3 Pro in the model selector.

Google AI

169,293 次观看 • 8 个月前

With this new release, Gemini applies advanced reasoning to make sense of complex, changing information in real-time. 3.1 Pro can utilize advanced reasoning to bridge the gap between complex APIs and user-friendly design. For example, Gemini built a live aerospace dashboard, taking live telemetry from a public API and designing a front end to track the Space Station's orbit.

With this new release, Gemini applies advanced reasoning to make sense of complex, changing information in real-time. 3.1 Pro can utilize advanced reasoning to bridge the gap between complex APIs and user-friendly design. For example, Gemini built a live aerospace dashboard, taking live telemetry from a public API and designing a front end to track the Space Station's orbit.

Google Gemini

106,360 次观看 • 5 个月前

Sentient took AI Startup of the Year at the Minsky Awards 🙌 ROMA Search achieved an industry-leading 81.7% accuracy, significantly surpassing competitors on the FRAMES benchmark for multi-step reasoning. On the SEAL-0 benchmark for complex, multi-source reasoning, ROMA with 45.6% accuracy more than doubled the score of Gemini 2.5 Pro (19.8%). ROMA is available on GitHub, try it out today👇

Sentient took AI Startup of the Year at the Minsky Awards 🙌 ROMA Search achieved an industry-leading 81.7% accuracy, significantly surpassing competitors on the FRAMES benchmark for multi-step reasoning. On the SEAL-0 benchmark for complex, multi-source reasoning, ROMA with 45.6% accuracy more than doubled the score of Gemini 2.5 Pro (19.8%). ROMA is available on GitHub, try it out today👇

Sentient

482,992 次观看 • 10 个月前

AI Mode is rolling out to everyone in the US. It’s a total reimagining of Search with more advanced reasoning so you can ask longer, complex queries. AI Overviews are now used by 1.5B people a month, in 200+ countries and territories. And Gemini 2.5 is coming to both this week.

AI Mode is rolling out to everyone in the US. It’s a total reimagining of Search with more advanced reasoning so you can ask longer, complex queries. AI Overviews are now used by 1.5B people a month, in 200+ countries and territories. And Gemini 2.5 is coming to both this week.

Sundar Pichai

61,207 次观看 • 1 年前

Bolt's agent now runs Claude 3.7 Sonnet—a huge upgrade for coding, design & debugging! But that’s not all… We’re introducing Dynamic Reasoning, a first-of-its-kind feature that lets the AI choose how hard to “think” when tackling complex problems.

Bolt's agent now runs Claude 3.7 Sonnet—a huge upgrade for coding, design & debugging! But that’s not all… We’re introducing Dynamic Reasoning, a first-of-its-kind feature that lets the AI choose how hard to “think” when tackling complex problems.

bolt.new

70,684 次观看 • 1 年前

MICROSOFT: "One of Grok 4’s headline features is its first-principles reasoning ability – essentially, the model tries to “think” like a scientist or detective, breaking problems down step by step."

MICROSOFT: "One of Grok 4’s headline features is its first-principles reasoning ability – essentially, the model tries to “think” like a scientist or detective, breaking problems down step by step."

DogeDesigner

19,246 次观看 • 10 个月前

🧠 Chat with Reasoning A few days ago the DeepSeek team released a LLM model with reasoning in various sizes. This we show is an example of 1bl that can run on machines with low GPU power like a mobile, but have enough power to answer complex questions. With these advanced models it is possible to link it with #IoT equipment to control information and use it in advanced control environments. All this under Open Source models and decentralized networks such as #Neurai #XNA $XNA #DeepSeek #Reasoning #AIchat

🧠 Chat with Reasoning A few days ago the DeepSeek team released a LLM model with reasoning in various sizes. This we show is an example of 1bl that can run on machines with low GPU power like a mobile, but have enough power to answer complex questions. With these advanced models it is possible to link it with #IoT equipment to control information and use it in advanced control environments. All this under Open Source models and decentralized networks such as #Neurai #XNA $XNA #DeepSeek #Reasoning #AIchat

NeurAI Project / XNA

17,691 次观看 • 1 年前

Today at #GoogleIO, we’re demonstrating how our partnership with Google Cloud is redefining the landscape of Intelligent Content Management. Leveraging the advanced capabilities of Google’s Gemini 2.5 Pro, our new AI Agent with enhanced data extraction transforms complex, unstructured data from various sources—including scanned PDFs, images, and diverse forms—into structured, actionable insights. Check out the demo below to watch Gemini 2.5 Pro and Box AI agents in action. With the Box AI Agent powered by Gemini 2.5 Pro you can: ⭐ Automatically, and instantly, classify and organize documents ⭐ Extract accurate insights and information from documents and images ⭐ Enable workflow automation by leveraging structured and easily retrievable content Our partnership with Google is not just about enhancing current capabilities; it’s about unlocking the full potential of AI, and models like Gemini 2.5 Pro, for intricate enterprise use cases.

Today at #GoogleIO, we’re demonstrating how our partnership with Google Cloud is redefining the landscape of Intelligent Content Management. Leveraging the advanced capabilities of Google’s Gemini 2.5 Pro, our new AI Agent with enhanced data extraction transforms complex, unstructured data from various sources—including scanned PDFs, images, and diverse forms—into structured, actionable insights. Check out the demo below to watch Gemini 2.5 Pro and Box AI agents in action. With the Box AI Agent powered by Gemini 2.5 Pro you can: ⭐ Automatically, and instantly, classify and organize documents ⭐ Extract accurate insights and information from documents and images ⭐ Enable workflow automation by leveraging structured and easily retrievable content Our partnership with Google is not just about enhancing current capabilities; it’s about unlocking the full potential of AI, and models like Gemini 2.5 Pro, for intricate enterprise use cases.

Box

10,565 次观看 • 1 年前

Today, we’re releasing a significant upgrade to our specialized reasoning mode, Gemini 3 Deep Think. Deep Think is built to drive practical applications, enabling researchers to interpret complex data and engineers to model physical systems through code. With the updated Deep Think, you can turn a sketch into a 3D-printable reality. Deep Think analyzes the drawing, builds the complex shape, and generates a file so you can create the physical object with 3D printing. This is rolling out now to Google AI Ultra subscribers. Select the "Deep Think" option in the tools menu to get started. Learn more here:

Today, we’re releasing a significant upgrade to our specialized reasoning mode, Gemini 3 Deep Think. Deep Think is built to drive practical applications, enabling researchers to interpret complex data and engineers to model physical systems through code. With the updated Deep Think, you can turn a sketch into a 3D-printable reality. Deep Think analyzes the drawing, builds the complex shape, and generates a file so you can create the physical object with 3D printing. This is rolling out now to Google AI Ultra subscribers. Select the "Deep Think" option in the tools menu to get started. Learn more here:

Google Gemini

5,055,337 次观看 • 6 个月前

Introducing Axiom’s discovery team, led by François Charton: We build models that will create novel constructions, map problems into solutions and intuitions, and learn the structure of entire mathematical worlds. Built to attack hard open problems, one at a time.

Introducing Axiom’s discovery team, led by François Charton: We build models that will create novel constructions, map problems into solutions and intuitions, and learn the structure of entire mathematical worlds. Built to attack hard open problems, one at a time.

Carina Hong

297,565 次观看 • 8 个月前

5⃣ Deep Research is even smarter, rolling out for everyone to try. Today, we're giving Gemini users the ability to try Deep Research, your personal AI research assistant, at no cost. 🧠✨ Gemini Deep Research has been upgraded to 2.0 Flash Thinking Experimental. With this model’s advanced reasoning, Gemini is even better at all stages of research, from planning to delivering even more insightful and detailed reports. Gemini Advanced and qualifying enterprise users have expanded access to Deep Research to save even more time on their most complex projects. Try Deep Research:

5⃣ Deep Research is even smarter, rolling out for everyone to try. Today, we're giving Gemini users the ability to try Deep Research, your personal AI research assistant, at no cost. 🧠✨ Gemini Deep Research has been upgraded to 2.0 Flash Thinking Experimental. With this model’s advanced reasoning, Gemini is even better at all stages of research, from planning to delivering even more insightful and detailed reports. Gemini Advanced and qualifying enterprise users have expanded access to Deep Research to save even more time on their most complex projects. Try Deep Research:

Google Gemini

258,461 次观看 • 1 年前

This is the first implementation of Grok voice API by xAI on a robot thanks to atariorbit (not perfect yet)! Feels like this could unlock some fun new use cases for robotics agents given that it ranks #1 on Big Bench Audio, the leading audio reasoning benchmark that measures voice agents’ capabilities to solve complex problems!

This is the first implementation of Grok voice API by xAI on a robot thanks to atariorbit (not perfect yet)! Feels like this could unlock some fun new use cases for robotics agents given that it ranks #1 on Big Bench Audio, the leading audio reasoning benchmark that measures voice agents’ capabilities to solve complex problems!

clem 🤗

45,390 次观看 • 7 个月前

✨ What makes Gemini 2.5 Pro stand out? In this Release Notes episode, Sr. Product Manager Logan Kilpatrick and Gemini Product Lead Tulsee Doshi break down its reasoning, coding, and multimodal strengths, plus a 1M token long context. ↓ Timecodes: 1:05 Gemini 2.5 launch overview 3:19 Academic evals vs. vibe checks 6:19 The jump to 2.5 7:51 Coordinating cross-stack improvements 11:48 Role of pre/post-training vs. test-time compute 13:21 Shipping Gemini 2.5 15:29 Embedded safety process 17:28 Multimodal reasoning with Gemini 2.5 18:55 Benchmark deep dive 22:07 What’s next for Gemini 24:49 Dynamic thinking in Gemini 2.5 25:37 The team effort behind the launch

✨ What makes Gemini 2.5 Pro stand out? In this Release Notes episode, Sr. Product Manager Logan Kilpatrick and Gemini Product Lead Tulsee Doshi break down its reasoning, coding, and multimodal strengths, plus a 1M token long context. ↓ Timecodes: 1:05 Gemini 2.5 launch overview 3:19 Academic evals vs. vibe checks 6:19 The jump to 2.5 7:51 Coordinating cross-stack improvements 11:48 Role of pre/post-training vs. test-time compute 13:21 Shipping Gemini 2.5 15:29 Embedded safety process 17:28 Multimodal reasoning with Gemini 2.5 18:55 Benchmark deep dive 22:07 What’s next for Gemini 24:49 Dynamic thinking in Gemini 2.5 25:37 The team effort behind the launch

Google AI Developers

228,165 次观看 • 1 年前