Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

🚀Introducing VisualWebBench: A Comprehensive Benchmark for Multimodal Web Page Understanding and Grounding. 🤔What's this all about? Why this benchmark? > Back in Nov 2023, when we released MMMU ( a comprehensive multimodal understanding benchmark, we received feedback that it included very few UI screenshots. Considering the growing importance of... UI understanding, especially with the rise of powerful agents like Devin ( which is built on the strong vision capability of #GPT4, we recognized the need for a benchmark focused on UI screenshot understanding.📸👀 > Multimodal #LLMs have significantly boosted web agents' performance on benchmarks like Mind2Web and WebArena. For instance, the SeeAct agent ( showcases the power of integrating vision into web agents. However, these benchmarks primarily evaluate the end-to-end task execution ability of web agents rather than their understanding of web pages. 🌉 Bridging the Gap with VisualWebBench > To provide a comprehensive evaluation of multimodal LLMs' web page understanding capabilities, we introduce VisualWebBench. Our benchmark spans 139 websites 🌐 across 12 domains 🏷️ and 87 sub-domains 🔍, ensuring a diverse and representative dataset. It assesses MLLMs at three levels: website-level, element-level, and action-level 📊, and encompasses seven tasks designed to evaluate understanding, OCR, grounding, and reasoning abilities 🧠💡. 😮 Surprising Findings > 🎉 Open-source models are catching up: Even though closed-source MLLMs are still leading the leaderboard, we are happy to see open-source models like LLaVA 1.6 34B achieve comparable performance to Gemini Pro. > 🧠 Grounding ability, crucial for developing MLLM-based web applications, is a weakness for most MLLMs. > 🖼️ Importance of Image Resolution: The limited image resolution handling capabilities of most open-source MLLMs restrict their utility in web scenarios, where rich text and elements are prevalent. > 🧱 Relatively strong correlation with general understanding benchmarks like MMMU but weak correlation with web agent benchmarks like Mind2Web. Web agent benchmarks primarily evaluate the end-to-end task execution ability of web agents, which involves a series of actions to accomplish a goal. In contrast, VisualWebBench emphasizes evaluating the foundational skills of MLLMs such as understanding and grounding web page elements. 💡Fun Fact > Claude Sonnet is better than Opus on our benchmark :) 🎓 Conclusion > VisualWebBench serves as a valuable resource for the community, driving research and development in the field of multimodal web page understanding and grounding. As MLLMs continue to evolve and improve, we look forward to seeing new applications and breakthroughs. We believe that our benchmark will contribute to the development of more powerful MLLMs in the web domain, ultimately leading to a more intuitive and efficient user experience on the web. Kudos to the student leads Junpeng Liu Yifan Song and the team Bill Yuchen Lin, Wai Lam, Graham Neubig, Yuanzhi Li! 👏 Check out more details in the Junpeng's thread👇show more

Xiang Yue

6,577 subscribers

56,696 Aufrufe • vor 2 Jahren •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

6 Kommentare

Profilbild von Li Bo

Li Bovor 2 Jahren

Congrats! May I know if it's possible to PR this dataset into lmms-eval?

Profilbild von Xiang Yue@ICLR2025🇸🇬

Xiang Yue@ICLR2025🇸🇬vor 2 Jahren

Sure! Happy to do that!

Profilbild von Travala.com 🏨 ✈️

Travala.com 🏨 ✈️vor 4 Jahren

This one's for the Crypto Whales ⬇️ NEW exclusive & luxurious experiences have been added to Apply today & #TravelUnlikeAnyOther with your favorite cryptocurrency.

Profilbild von Homer Quan

Homer Quanvor 2 Jahren

Nice work, this will be the foundational block of emerging “deep search” business model where Google is trying to pivot its search business into)

Profilbild von Janak Agrawal

Janak Agrawalvor 2 Jahren

This benchmark is amazing!! Does this benchmark reference Gemini 1 Pro or Gemini 1.5 Pro?

Profilbild von Xiang Yue@ICLR2025🇸🇬

Xiang Yue@ICLR2025🇸🇬vor 2 Jahren

Thanks! We used Gemini 1 Pro.

Ähnliche Videos

Working on multimodal instruction tuning and finding it hard to scale? Building Web/GUI agents but data is too narrow? Introducing 🚀MultiUI: 7.3M multimodal instructions from 1M webpage UIs, offering diverse data to boost text-rich visual understanding. Key takeaways: 🌟WebUI-trained models show major gains in visual web understanding and agent tasks. 💻 🌟Models also generalize well to non-UI tasks like DocVQA/OCR. 📄 How it works: We generate multimodal instructions with a text LLM using structured text from webpage accessibility trees. We then pair them with UI screenshots, to train multimodal models. Homepage: Paper: Dataset: Model: Congrats to the student lead Junpeng Liu and the team Tianyue Ou Yifan Song Yuxiao Qu Chenyan Xiong Wenhu Chen Graham Neubig ! More details are in the following threads ⬇️

Working on multimodal instruction tuning and finding it hard to scale? Building Web/GUI agents but data is too narrow? Introducing 🚀MultiUI: 7.3M multimodal instructions from 1M webpage UIs, offering diverse data to boost text-rich visual understanding. Key takeaways: 🌟WebUI-trained models show major gains in visual web understanding and agent tasks. 💻 🌟Models also generalize well to non-UI tasks like DocVQA/OCR. 📄 How it works: We generate multimodal instructions with a text LLM using structured text from webpage accessibility trees. We then pair them with UI screenshots, to train multimodal models. Homepage: Paper: Dataset: Model: Congrats to the student lead Junpeng Liu and the team Tianyue Ou Yifan Song Yuxiao Qu Chenyan Xiong Wenhu Chen Graham Neubig ! More details are in the following threads ⬇️

Xiang Yue

57,729 Aufrufe • vor 1 Jahr

New Short Course: Building AI Browser Agents! Learn how to build AI agents that interact and take actions on websites in this course, created in partnership with and taught by and @namangarg0, Co-founders of AGI Inc. AI browser agents can log into websites, fill out forms, click through web pages, or even place orders online for you. They use both visual information, like screenshots, and structural data, like the HTML or Document Object Model (DOM) of a web page, to reason and take action. With the complexity of webpages and multiple possible actions at each step, it can be challenging for an AI browser agent to complete an assigned task. Because these agents run long action sequences, a single error—like clicking the wrong button or misreading a field—can lead to unexpected outcomes or errors that compound over time. In this course, you'll understand how autonomous web agents work, their current limitations, and how AgentQ enables them to improve through self-correction. In detail, you'll: - Learn what web agents are, how they automate tasks online, their architecture, key components, limitations, and an overview of their decision-making strategies. - Build a web agent that can scrape website and return course recommendations in a structured output format. - Build an autonomous web agent that can execute multiple tasks, such as finding and summarizing webpages, filling out a form, and signing up for a newsletter. - Explore AgentQ, a framework that enables agents to self-correct by combining Monte Carlo Tree Search (MCTS), a self-critique mechanism for continuous improvement, and Direct Preference Optimization (DPO). - Deep dive into MCTS, learn how it finds an effective path, illustrated by an example of Gridworld animation, and use AgentQ to complete web tasks. - Understand AI agents' current state and future directions—including key factors shaping their evolution, such as hardware, algorithm innovation, and data availability. By the end of this course, you will have hands-on experience building browser agents and a deeper understanding of how to make them more robust and reliable. Please sign up here:

New Short Course: Building AI Browser Agents! Learn how to build AI agents that interact and take actions on websites in this course, created in partnership with and taught by and @namangarg0, Co-founders of AGI Inc. AI browser agents can log into websites, fill out forms, click through web pages, or even place orders online for you. They use both visual information, like screenshots, and structural data, like the HTML or Document Object Model (DOM) of a web page, to reason and take action. With the complexity of webpages and multiple possible actions at each step, it can be challenging for an AI browser agent to complete an assigned task. Because these agents run long action sequences, a single error—like clicking the wrong button or misreading a field—can lead to unexpected outcomes or errors that compound over time. In this course, you'll understand how autonomous web agents work, their current limitations, and how AgentQ enables them to improve through self-correction. In detail, you'll: - Learn what web agents are, how they automate tasks online, their architecture, key components, limitations, and an overview of their decision-making strategies. - Build a web agent that can scrape website and return course recommendations in a structured output format. - Build an autonomous web agent that can execute multiple tasks, such as finding and summarizing webpages, filling out a form, and signing up for a newsletter. - Explore AgentQ, a framework that enables agents to self-correct by combining Monte Carlo Tree Search (MCTS), a self-critique mechanism for continuous improvement, and Direct Preference Optimization (DPO). - Deep dive into MCTS, learn how it finds an effective path, illustrated by an example of Gridworld animation, and use AgentQ to complete web tasks. - Understand AI agents' current state and future directions—including key factors shaping their evolution, such as hardware, algorithm innovation, and data availability. By the end of this course, you will have hands-on experience building browser agents and a deeper understanding of how to make them more robust and reliable. Please sign up here:

Andrew Ng

186,182 Aufrufe • vor 1 Jahr

The web was never meant to be flattened into text. Yet most web RAG systems start by parsing HTML --- a complex and lossy process. 🔥 Introducing PixelRAG: the first RAG system that retrieves and reads 30M+ web pages as pixels. Instead of extracting text, PixelRAG retrieves screenshots and lets a VLM read them directly. PixelRAG not only preserves visual information, but also outperforms text-based RAG on text-only QA benchmarks by +18.1%. Why? (1) HTML-to-text conversion often discards layout, structure, tables, and other useful signals. (2) We continued pretraining a VLM on web page screenshots and turned it into a surprisingly strong visual retriever. (3) Recent VLMs are remarkably good at understanding web pages, often with better accuracy and token efficiency than text-only pipelines. Takeaway: HTML parsing may be one of the biggest self-inflicted bottlenecks in web RAG. Demo below 👇 Code: Paper: Playground:

The web was never meant to be flattened into text. Yet most web RAG systems start by parsing HTML --- a complex and lossy process. 🔥 Introducing PixelRAG: the first RAG system that retrieves and reads 30M+ web pages as pixels. Instead of extracting text, PixelRAG retrieves screenshots and lets a VLM read them directly. PixelRAG not only preserves visual information, but also outperforms text-based RAG on text-only QA benchmarks by +18.1%. Why? (1) HTML-to-text conversion often discards layout, structure, tables, and other useful signals. (2) We continued pretraining a VLM on web page screenshots and turned it into a surprisingly strong visual retriever. (3) Recent VLMs are remarkably good at understanding web pages, often with better accuracy and token efficiency than text-only pipelines. Takeaway: HTML parsing may be one of the biggest self-inflicted bottlenecks in web RAG. Demo below 👇 Code: Paper: Playground:

Yichuan Wang

91,784 Aufrufe • vor 3 Monaten

How well do today’s frontier models handle long-horizon, multi-step web agent tasks, such as identifying the top 25 U.S. CS PhD programs with ML/AI faculty likely accepting students and compiling the results into a structured sheet? Check out our new work on Odysseys: Benchmarking Web Agents on Realistic Long Horizon Tasks Paper: Leaderboard: We introduce Odysseys, a benchmark of 200 long-horizon tasks derived from real browsing sessions and evaluated on the live Internet. We show that binary pass/fail is inadequate in this setting and propose rubric-based evaluation, which better aligns with human judgment and provides more informative signals. Across leading models, the best achieves only 44.5% success, leaving substantial headroom. We further introduce a Trajectory Efficiency metric (rubric score per step) and find efficiency remains extremely low (1.15%), highlighting a key bottleneck. Odysseys provides a realistic benchmark for measuring progress toward web agents capable of sustained, efficient, real-world operation. See a more detailed thread by Jing Yu Koh.

How well do today’s frontier models handle long-horizon, multi-step web agent tasks, such as identifying the top 25 U.S. CS PhD programs with ML/AI faculty likely accepting students and compiling the results into a structured sheet? Check out our new work on Odysseys: Benchmarking Web Agents on Realistic Long Horizon Tasks Paper: Leaderboard: We introduce Odysseys, a benchmark of 200 long-horizon tasks derived from real browsing sessions and evaluated on the live Internet. We show that binary pass/fail is inadequate in this setting and propose rubric-based evaluation, which better aligns with human judgment and provides more informative signals. Across leading models, the best achieves only 44.5% success, leaving substantial headroom. We further introduce a Trajectory Efficiency metric (rubric score per step) and find efficiency remains extremely low (1.15%), highlighting a key bottleneck. Odysseys provides a realistic benchmark for measuring progress toward web agents capable of sustained, efficient, real-world operation. See a more detailed thread by Jing Yu Koh.

Russ Salakhutdinov

22,518 Aufrufe • vor 4 Monaten

Achieve SOTA on multimodal, complex tasks with minimal supervision! Introducing In-Context Abstraction Learning (ICAL): a method that extracts insights from visual demos and human feedback to continually teach multimodal LLM agents 🚀🤖🧠 ICAL learns from sub-optimal demonstrations and human feedback by transforming noisy trajectories into optimized, annotated examples that improve multimodal decision-making in new domains. ICAL is the first to show state-of-the-art performance across three challenging domains: we obtain a 34% relative improvement on TEACh household robot instruction following from dialogue, a 59% relative improvement on Visual Web Arena for autonomous visual web agents, and beat few-shot GPT-4V and supervised baselines on Ego4D action forecasting.

Achieve SOTA on multimodal, complex tasks with minimal supervision! Introducing In-Context Abstraction Learning (ICAL): a method that extracts insights from visual demos and human feedback to continually teach multimodal LLM agents 🚀🤖🧠 ICAL learns from sub-optimal demonstrations and human feedback by transforming noisy trajectories into optimized, annotated examples that improve multimodal decision-making in new domains. ICAL is the first to show state-of-the-art performance across three challenging domains: we obtain a 34% relative improvement on TEACh household robot instruction following from dialogue, a 59% relative improvement on Visual Web Arena for autonomous visual web agents, and beat few-shot GPT-4V and supervised baselines on Ego4D action forecasting.

Gabriel Sarch

19,340 Aufrufe • vor 2 Jahren

LLaVA-3D A Simple yet Effective Pathway to Empowering LMMs with 3D-awareness Recent advancements in Large Multimodal Models (LMMs) have greatly enhanced their proficiency in 2D visual understanding tasks, enabling them to effectively process and understand images and videos. However, the development of LMMs with 3D-awareness for 3D scene understanding has been hindered by the lack of large-scale 3D vision-language datasets and powerful 3D encoders. In this paper, we introduce a simple yet effective framework called LLaVA-3D. Leveraging the strong 2D understanding priors from LLaVA, our LLaVA-3D efficiently adapts LLaVA for 3D scene understanding without compromising 2D understanding capabilities. To achieve this, we employ a simple yet effective representation, 3D Patch, which connects 2D CLIP patch features with their corresponding positions in 3D space. By integrating the 3D Patches into 2D LMMs and employing joint 2D and 3D vision-language instruction tuning, we establish a unified architecture for both 2D image understanding and 3D scene understanding. Experimental results show that LLaVA-3D converges 3.5x faster than existing 3D LMMs when trained on 3D vision-language datasets. Moreover, LLaVA-3D not only achieves state-of-the-art performance across various 3D tasks but also maintains comparable 2D image understanding and vision-language conversation capabilities with LLaVA.

LLaVA-3D A Simple yet Effective Pathway to Empowering LMMs with 3D-awareness Recent advancements in Large Multimodal Models (LMMs) have greatly enhanced their proficiency in 2D visual understanding tasks, enabling them to effectively process and understand images and videos. However, the development of LMMs with 3D-awareness for 3D scene understanding has been hindered by the lack of large-scale 3D vision-language datasets and powerful 3D encoders. In this paper, we introduce a simple yet effective framework called LLaVA-3D. Leveraging the strong 2D understanding priors from LLaVA, our LLaVA-3D efficiently adapts LLaVA for 3D scene understanding without compromising 2D understanding capabilities. To achieve this, we employ a simple yet effective representation, 3D Patch, which connects 2D CLIP patch features with their corresponding positions in 3D space. By integrating the 3D Patches into 2D LMMs and employing joint 2D and 3D vision-language instruction tuning, we establish a unified architecture for both 2D image understanding and 3D scene understanding. Experimental results show that LLaVA-3D converges 3.5x faster than existing 3D LMMs when trained on 3D vision-language datasets. Moreover, LLaVA-3D not only achieves state-of-the-art performance across various 3D tasks but also maintains comparable 2D image understanding and vision-language conversation capabilities with LLaVA.

AK

41,747 Aufrufe • vor 2 Jahren

LLM agents have demonstrated promise in their ability to automate computer tasks, but face challenges with multi-step reasoning and planning. Towards addressing this, we propose an inference-time tree search algorithm for LLM agents to explicitly perform exploration and multi-step planning in interactive web environments. It is the first tree search algorithm for LLM agents that shows effectiveness on realistic and complex web environments: on the challenging VisualWebArena benchmark, applying our search algorithm on top of a GPT-4o agent yields a 39.7% relative increase in success rate compared to the same baseline without search, setting a state-of-the-art success rate of 26.4%. On WebArena, search also yields a 28.0% relative improvement over a baseline agent, setting a competitive success rate of 19.2%.

LLM agents have demonstrated promise in their ability to automate computer tasks, but face challenges with multi-step reasoning and planning. Towards addressing this, we propose an inference-time tree search algorithm for LLM agents to explicitly perform exploration and multi-step planning in interactive web environments. It is the first tree search algorithm for LLM agents that shows effectiveness on realistic and complex web environments: on the challenging VisualWebArena benchmark, applying our search algorithm on top of a GPT-4o agent yields a 39.7% relative increase in success rate compared to the same baseline without search, setting a state-of-the-art success rate of 26.4%. On WebArena, search also yields a 28.0% relative improvement over a baseline agent, setting a competitive success rate of 19.2%.

Jing Yu Koh

124,894 Aufrufe • vor 2 Jahren

Today we are announcing Keenable AI, an AI-native index of the best human knowledge we have, starting with the open web. AI seems to know just about everything until you ask it about something you know deeply. The answer isn’t wrong, but it's just very average. We started Keenable to fix this problem: every model and every agent should be able to query, reason over, and continuously learn from the living web. Backed by a $26M Seed from Accel and Conviction. Built by the team that took on Google at Yandex Search, with researchers and engineers from Amazon AGI, X, and Perplexity. The Web Search API and Web Query Language for AI are live now. Together they make the web queryable at AI scale. Both are free until end of September. Seven days of benchmarks, stories, and launches ahead. Let your agents search.

Today we are announcing Keenable AI, an AI-native index of the best human knowledge we have, starting with the open web. AI seems to know just about everything until you ask it about something you know deeply. The answer isn’t wrong, but it's just very average. We started Keenable to fix this problem: every model and every agent should be able to query, reason over, and continuously learn from the living web. Backed by a $26M Seed from Accel and Conviction. Built by the team that took on Google at Yandex Search, with researchers and engineers from Amazon AGI, X, and Perplexity. The Web Search API and Web Query Language for AI are live now. Together they make the web queryable at AI scale. Both are free until end of September. Seven days of benchmarks, stories, and launches ahead. Let your agents search.

Andrey Styskin

550,073 Aufrufe • vor 25 Tagen

🚨 THIS IS ACTUALLY INSANE Your AI agent can have access to the web. But if it can't reliably read what’s actually on the page, that access is almost useless. We looked at Firecrawl as the web layer for AI agents and the numbers are hard to ignore. The setup is simple: Give it a URL, search query, or website. Firecrawl handles the ugly part — crawling, scraping, rendering, extracting, and turning web content into something an AI model can actually use. The headline numbers: → 173,000+ GitHub stars → Search, scrape and interact with the web at scale → Supports web pages, PDFs, DOCX and other content → Structured data extraction for AI workflows → MCP support for connecting it directly to AI agents The workflow looks like this: Search → Scrape → Crawl → Extract → Feed the agent Three things stand out: 1. Scraping becomes an infrastructure layer Instead of maintaining your own pile of HTTP clients, parsers, browser automation and retry logic, you can treat web access as an API. 2. Agents get more than raw HTML The goal isn't just downloading a webpage. It's turning messy web content into clean context that an LLM can reason over. 3. The same layer works across different agent workflows Research agents. RAG pipelines. AI search. Competitive intelligence. Web-data extraction. The interesting shift: AI agents don't just need better models. They need better access to the information those models are supposed to reason about. Firecrawl is building that layer. Save this repo.

🚨 THIS IS ACTUALLY INSANE Your AI agent can have access to the web. But if it can't reliably read what’s actually on the page, that access is almost useless. We looked at Firecrawl as the web layer for AI agents and the numbers are hard to ignore. The setup is simple: Give it a URL, search query, or website. Firecrawl handles the ugly part — crawling, scraping, rendering, extracting, and turning web content into something an AI model can actually use. The headline numbers: → 173,000+ GitHub stars → Search, scrape and interact with the web at scale → Supports web pages, PDFs, DOCX and other content → Structured data extraction for AI workflows → MCP support for connecting it directly to AI agents The workflow looks like this: Search → Scrape → Crawl → Extract → Feed the agent Three things stand out: 1. Scraping becomes an infrastructure layer Instead of maintaining your own pile of HTTP clients, parsers, browser automation and retry logic, you can treat web access as an API. 2. Agents get more than raw HTML The goal isn't just downloading a webpage. It's turning messy web content into clean context that an LLM can reason over. 3. The same layer works across different agent workflows Research agents. RAG pipelines. AI search. Competitive intelligence. Web-data extraction. The interesting shift: AI agents don't just need better models. They need better access to the information those models are supposed to reason about. Firecrawl is building that layer. Save this repo.

Vikas gupta

18,053 Aufrufe • vor 14 Tagen

"The future of AI is agentic. That includes browsers!" Imagine having an AI agent in your browser that can help you complete complex tasks, answer your questions, and streamline your workflow. Today I'm thrilled to share a sneak peek at Project Mariner, a cutting-edge research collaboration between Chrome and Google DeepMind, exploring the future of agentic AI within the browser! Building on the power of Gemini 2.0, Mariner envisions AI agents seamlessly guiding users through online tasks, streamlining workflows and enriching browsing experiences. Imagine having an intelligent co-pilot in your browser, anticipating your needs and proactively offering assistance. We're in the early stages of experimentation, focusing on core functionalities like understanding user intent, automating actions, and providing personalized recommendations. This prototype leverages Gemini's advanced natural language understanding and reasoning capabilities to interpret user requests, both typed and spoken. Mariner can then interact with web pages, retrieve information, and even perform actions like filling out forms or navigating to specific sites. For example, a user could simply ask "Find me a job near me," and Mariner would understand the request, navigate to a relevant job search site, and tailor the search based on the user's location and preferences. This is just one example of how we're exploring Gemini 2.0's potential to unlock agentic experiences through a series of prototypes, including: 1. Agents with multimodal reasoning: Project Astra, our research prototype exploring the capabilities of a universal AI assistant, is enhanced by Gemini 2.0. 2. Agents that can help you accomplish complex tasks: Project Mariner itself focuses on the future of human-agent interaction within the browser. 3. Agents for developers: Jules is an experimental AI-powered coding agent that integrates directly into a GitHub workflow. 4. Agents applied across domains: We're exploring agents for navigating video games and even applying Gemini 2.0's spatial reasoning to robotics. We believe that integrating AI agents directly into the browser has the potential to revolutionize how we interact with the web. Project Mariner aims to make browsing more intuitive, efficient, and personalized. By understanding user context and proactively offering assistance, Mariner can simplify complex tasks, save users time, and empower them to achieve more online. This aligns perfectly with the vision of Gemini 2.0 to create more helpful and intuitive AI experiences. We’re currently testing Mariner with a small group of trusted users to gather feedback and refine the user experience. We believe that this technology holds immense potential to transform the way we browse and interact with information online.

"The future of AI is agentic. That includes browsers!" Imagine having an AI agent in your browser that can help you complete complex tasks, answer your questions, and streamline your workflow. Today I'm thrilled to share a sneak peek at Project Mariner, a cutting-edge research collaboration between Chrome and Google DeepMind, exploring the future of agentic AI within the browser! Building on the power of Gemini 2.0, Mariner envisions AI agents seamlessly guiding users through online tasks, streamlining workflows and enriching browsing experiences. Imagine having an intelligent co-pilot in your browser, anticipating your needs and proactively offering assistance. We're in the early stages of experimentation, focusing on core functionalities like understanding user intent, automating actions, and providing personalized recommendations. This prototype leverages Gemini's advanced natural language understanding and reasoning capabilities to interpret user requests, both typed and spoken. Mariner can then interact with web pages, retrieve information, and even perform actions like filling out forms or navigating to specific sites. For example, a user could simply ask "Find me a job near me," and Mariner would understand the request, navigate to a relevant job search site, and tailor the search based on the user's location and preferences. This is just one example of how we're exploring Gemini 2.0's potential to unlock agentic experiences through a series of prototypes, including: 1. Agents with multimodal reasoning: Project Astra, our research prototype exploring the capabilities of a universal AI assistant, is enhanced by Gemini 2.0. 2. Agents that can help you accomplish complex tasks: Project Mariner itself focuses on the future of human-agent interaction within the browser. 3. Agents for developers: Jules is an experimental AI-powered coding agent that integrates directly into a GitHub workflow. 4. Agents applied across domains: We're exploring agents for navigating video games and even applying Gemini 2.0's spatial reasoning to robotics. We believe that integrating AI agents directly into the browser has the potential to revolutionize how we interact with the web. Project Mariner aims to make browsing more intuitive, efficient, and personalized. By understanding user context and proactively offering assistance, Mariner can simplify complex tasks, save users time, and empower them to achieve more online. This aligns perfectly with the vision of Gemini 2.0 to create more helpful and intuitive AI experiences. We’re currently testing Mariner with a small group of trusted users to gather feedback and refine the user experience. We believe that this technology holds immense potential to transform the way we browse and interact with information online.

Addy Osmani

29,518 Aufrufe • vor 1 Jahr

Earlier this week at GTC, we announced our partnership with Nvidia. We will work with Nvidia to build strong, American open-source models that are at the frontier of scientific reasoning. These models will be essential for the US to compete with China on science in the coming decades. Jensen is committing to spend tens of billions of dollars developing open-source models, and we are excited to be a partner with them in figuring out how to benchmark, train and use those agents to accelerate scientific research. We have already open-sourced some of the work we have done with them, and are looking forward to open-sourcing more. There are few things today that are more important. See our blog post below, and watch the video to learn more, narrated by the man himself.

Earlier this week at GTC, we announced our partnership with Nvidia. We will work with Nvidia to build strong, American open-source models that are at the frontier of scientific reasoning. These models will be essential for the US to compete with China on science in the coming decades. Jensen is committing to spend tens of billions of dollars developing open-source models, and we are excited to be a partner with them in figuring out how to benchmark, train and use those agents to accelerate scientific research. We have already open-sourced some of the work we have done with them, and are looking forward to open-sourcing more. There are few things today that are more important. See our blog post below, and watch the video to learn more, narrated by the man himself.

Sam Rodriques

24,338 Aufrufe • vor 6 Monaten

We’re officially a YOUnicorn! 🦄 Excited to share that You.com just raised $100M Series C at a $1.5B valuation, led by Cox Enterprises. We’ve been heads down building the search infrastructure for the AI and agent future. Soon there will be more AI agents using the web than humans, but today's search wasn't built for this. Agents need deep, contextual information from both public web and internal private data to make real decisions. Our web search API delivers the most up-to-date, accurate, and fastest search results for LLMs and agents. Real benchmarks show we consistently outperform the competition on accuracy and speed while staying cost-effective. Today, we’re answering over 1 billion queries every month across enterprises like DuckDuckGo, Windsurf, Harvey, and many others. There is no other startup of recent years at this scale. Grateful to our wonderful employees, customers, partners, and investors who believe in our vision. The agentic era is here - let's build it together. Try our LLM search APIs.

We’re officially a YOUnicorn! 🦄 Excited to share that You.com just raised $100M Series C at a $1.5B valuation, led by Cox Enterprises. We’ve been heads down building the search infrastructure for the AI and agent future. Soon there will be more AI agents using the web than humans, but today's search wasn't built for this. Agents need deep, contextual information from both public web and internal private data to make real decisions. Our web search API delivers the most up-to-date, accurate, and fastest search results for LLMs and agents. Real benchmarks show we consistently outperform the competition on accuracy and speed while staying cost-effective. Today, we’re answering over 1 billion queries every month across enterprises like DuckDuckGo, Windsurf, Harvey, and many others. There is no other startup of recent years at this scale. Grateful to our wonderful employees, customers, partners, and investors who believe in our vision. The agentic era is here - let's build it together. Try our LLM search APIs.

Richard Socher

63,660 Aufrufe • vor 1 Jahr

We’re launching Optima. Now anyone can create a custom benchmark for their use case, leveraging Artificial Analysis’ leading research and platform Building and running benchmarks is difficult. We have distilled Artificial Analysis’ research and experience developing benchmarks into Optima, a new platform for benchmarking models on your own workloads and comparing performance, speed and cost efficiency. Optima allows you to find the best model for your task, or an equally performant alternative to your current setup at 10x lower cost or time per task. We’ve integrated Artificial Analysis' research and experience in benchmarks across the Optima workflow: ➤ Build benchmarks based on your own data and use cases: There are three ways to build a benchmark with Optima. Upload an existing evaluation dataset from your own files or Hugging Face, or import agent traces from platforms including Arize AI, Braintrust and langfuse.com. Install the Optima skill to build a benchmark using context from your coding environment and previous sessions. Or simply describe your use case and provide example inputs and outputs, and Optima will build the benchmark for you ➤ Run across the latest models: Run the same benchmark across leading models in a single click, and keep your leaderboard up to date as soon as new models are released ➤ Bring Artificial Analysis grading to your own benchmark: Evaluate responses against objective rubric criteria or using the same pairwise judging approach used for Artificial Analysis benchmarks including GDPval-AA and AA-Briefcase. For pairwise judging, select your preferred responses from a sample and Optima uses those preferences to rank models across your test set ➤ Compare performance, cost and time efficiency: Optima measures more than model performance. Cost per Task and Time per Task are tracked alongside benchmark scores, with category-level results and support for custom metrics, allowing you to compare the tradeoffs between models for your specific use case Ahead of launch, here are examples questions our beta testers answered with Optima: ➤ Which model can save me 10x the cost without a meaningful decrease in quality for my finance & accounting agent? ➤ Which model best matches the writing style of lawyers for my legal agent? ➤ Which model can best identify different elements in my custom image dataset? Optima is available today. Build your own benchmark at

We’re launching Optima. Now anyone can create a custom benchmark for their use case, leveraging Artificial Analysis’ leading research and platform Building and running benchmarks is difficult. We have distilled Artificial Analysis’ research and experience developing benchmarks into Optima, a new platform for benchmarking models on your own workloads and comparing performance, speed and cost efficiency. Optima allows you to find the best model for your task, or an equally performant alternative to your current setup at 10x lower cost or time per task. We’ve integrated Artificial Analysis' research and experience in benchmarks across the Optima workflow: ➤ Build benchmarks based on your own data and use cases: There are three ways to build a benchmark with Optima. Upload an existing evaluation dataset from your own files or Hugging Face, or import agent traces from platforms including Arize AI, Braintrust and langfuse.com. Install the Optima skill to build a benchmark using context from your coding environment and previous sessions. Or simply describe your use case and provide example inputs and outputs, and Optima will build the benchmark for you ➤ Run across the latest models: Run the same benchmark across leading models in a single click, and keep your leaderboard up to date as soon as new models are released ➤ Bring Artificial Analysis grading to your own benchmark: Evaluate responses against objective rubric criteria or using the same pairwise judging approach used for Artificial Analysis benchmarks including GDPval-AA and AA-Briefcase. For pairwise judging, select your preferred responses from a sample and Optima uses those preferences to rank models across your test set ➤ Compare performance, cost and time efficiency: Optima measures more than model performance. Cost per Task and Time per Task are tracked alongside benchmark scores, with category-level results and support for custom metrics, allowing you to compare the tradeoffs between models for your specific use case Ahead of launch, here are examples questions our beta testers answered with Optima: ➤ Which model can save me 10x the cost without a meaningful decrease in quality for my finance & accounting agent? ➤ Which model best matches the writing style of lawyers for my legal agent? ➤ Which model can best identify different elements in my custom image dataset? Optima is available today. Build your own benchmark at

Artificial Analysis

132,623 Aufrufe • vor 1 Monat

Super excited to share 🧠MLGym 🦾 – the first Gym environment for AI Research Agents 🤖🔬 We introduce MLGym and MLGym-Bench, a new framework and benchmark for evaluating and developing LLM agents on AI research tasks. The key contributions of our work are: 🕹️ Enables the exploration of different training algorithms for AI Research Agents such as RL 🛠️ Provides a flexible evaluation framework that can accommodate different artifacts such as models, algorithms, or predictions 🤖 Allows researchers to evaluate any model without the need to develop a custom agentic harness 🎯 Introduces 13 diverse open-ended AI Research tasks for evaluating AI Research Agents on a wide range of domains such as computer vision, natural language processing, reinforcement learning, game theory, and logical reasoning. 📈 Proposes a new evaluation metric for AI Research Agents MLGym makes it easy to: 1) Add new tasks 2) Evaluate new models 3) Integrate new agents Check out a video of the MLGym Agent to see how it performs the full pipeline of idea generation💡, implementation 👩‍💻, experimentation 👩‍🔬, and iteration 🔄 to improve on ML tasks. Huge thanks to the exceptionally talented Deepak Nathani who led this work and to all the other amazing collaborators who made this possible 🙏🫶🚀

Super excited to share 🧠MLGym 🦾 – the first Gym environment for AI Research Agents 🤖🔬 We introduce MLGym and MLGym-Bench, a new framework and benchmark for evaluating and developing LLM agents on AI research tasks. The key contributions of our work are: 🕹️ Enables the exploration of different training algorithms for AI Research Agents such as RL 🛠️ Provides a flexible evaluation framework that can accommodate different artifacts such as models, algorithms, or predictions 🤖 Allows researchers to evaluate any model without the need to develop a custom agentic harness 🎯 Introduces 13 diverse open-ended AI Research tasks for evaluating AI Research Agents on a wide range of domains such as computer vision, natural language processing, reinforcement learning, game theory, and logical reasoning. 📈 Proposes a new evaluation metric for AI Research Agents MLGym makes it easy to: 1) Add new tasks 2) Evaluate new models 3) Integrate new agents Check out a video of the MLGym Agent to see how it performs the full pipeline of idea generation💡, implementation 👩‍💻, experimentation 👩‍🔬, and iteration 🔄 to improve on ML tasks. Huge thanks to the exceptionally talented Deepak Nathani who led this work and to all the other amazing collaborators who made this possible 🙏🫶🚀

Roberta Raileanu

105,157 Aufrufe • vor 1 Jahr

Destra Network| One-Click Decentralized Hosting - Upgrade: v0.2 Today, we are excited to launch version 0.2 of Destra Decentralized Web Hosting, which introduces multiple enhancements and upgrades over the existing v0.1. Here's a detailed breakdown of the updates and features in v0.2: - Enhanced UI/UX: Based on community feedback, we've made significant design improvements to the user interface and user experience. The UI is now more user-friendly and comprehensive compared to v0.1. - Live Deployment Previews: Users can now view and monitor websites they've deployed on Destra Decentralized Web Hosting in real-time. Additionally, there's functionality to check the health of deployments on the Destra Network. -Destra DNS Management: Version 0.2 introduces a UI for Destra DNS management, including the ability to manage DNS zone records. - ENS Domains Management UI: The v0.2 update includes a user interface for managing ENS domains, enhancing control and accessibility, the functionality will be going live in next performance upgrade. The video attached below demonstrates the process.

Destra Network| One-Click Decentralized Hosting - Upgrade: v0.2 Today, we are excited to launch version 0.2 of Destra Decentralized Web Hosting, which introduces multiple enhancements and upgrades over the existing v0.1. Here's a detailed breakdown of the updates and features in v0.2: - Enhanced UI/UX: Based on community feedback, we've made significant design improvements to the user interface and user experience. The UI is now more user-friendly and comprehensive compared to v0.1. - Live Deployment Previews: Users can now view and monitor websites they've deployed on Destra Decentralized Web Hosting in real-time. Additionally, there's functionality to check the health of deployments on the Destra Network. -Destra DNS Management: Version 0.2 introduces a UI for Destra DNS management, including the ability to manage DNS zone records. - ENS Domains Management UI: The v0.2 update includes a user interface for managing ENS domains, enhancing control and accessibility, the functionality will be going live in next performance upgrade. The video attached below demonstrates the process.

Destra Network

29,327 Aufrufe • vor 2 Jahren

Introducing ALE-Bench, ALE-Agent! Towards Automating Long-Horizon Algorithm Engineering for Hard Optimization Problems Blog: Paper: ALE-Bench is a coding benchmark primarily focused on hard optimization (NP-hard) problems. We developed this benchmark with AtCoder Inc., a leading coding contest platform company. What makes ALE-Bench unique is its focus on hard optimization problems that demand long-horizon and creative reasoning. It’s open-ended, in the sense that true optima are out of reach (NP-hard) and scores can continuously improve. We believe this benchmark has the potential to become one of the key benchmarks for reasoning and coding in the next generation. ALE-Agent is our end-to-end agent that we specifically designed for this challenging domain. In fact, our ALE-Agent has already built an impressive track record in the wild! In May 2025, our agent participated in a live AtCoder Heuristic Competition (AHC), alongside 1,000 other participants in real-time. AHC is considered to be one of the most challenging coding competitions in this domain. Our ALE-Agent achieved an impressive ranking of 21st out of 1,000 human participants in the competition (top 2%), marking a turning point for AI discovery of solutions to hard optimization problems with a wide spectrum of important real world applications such as logistics, routing, packing, factory production planning, power-grid balancing. We look forward to applying this technology to real industrial optimization opportunities. Building on the insights from this study, Sakana AI will continue to tackle the challenge of developing AI with even greater algorithm engineering capabilities. ALE-Bench Dataset: ALE-Bench Code: This research was conducted in collaboration with AtCoder Inc. (AtCoder). We are deeply grateful for their outstanding expertise and contributions in optimization and algorithms, which were invaluable in providing data, analyzing results, and enabling our AI agent’s participation in their contests.

Introducing ALE-Bench, ALE-Agent! Towards Automating Long-Horizon Algorithm Engineering for Hard Optimization Problems Blog: Paper: ALE-Bench is a coding benchmark primarily focused on hard optimization (NP-hard) problems. We developed this benchmark with AtCoder Inc., a leading coding contest platform company. What makes ALE-Bench unique is its focus on hard optimization problems that demand long-horizon and creative reasoning. It’s open-ended, in the sense that true optima are out of reach (NP-hard) and scores can continuously improve. We believe this benchmark has the potential to become one of the key benchmarks for reasoning and coding in the next generation. ALE-Agent is our end-to-end agent that we specifically designed for this challenging domain. In fact, our ALE-Agent has already built an impressive track record in the wild! In May 2025, our agent participated in a live AtCoder Heuristic Competition (AHC), alongside 1,000 other participants in real-time. AHC is considered to be one of the most challenging coding competitions in this domain. Our ALE-Agent achieved an impressive ranking of 21st out of 1,000 human participants in the competition (top 2%), marking a turning point for AI discovery of solutions to hard optimization problems with a wide spectrum of important real world applications such as logistics, routing, packing, factory production planning, power-grid balancing. We look forward to applying this technology to real industrial optimization opportunities. Building on the insights from this study, Sakana AI will continue to tackle the challenge of developing AI with even greater algorithm engineering capabilities. ALE-Bench Dataset: ALE-Bench Code: This research was conducted in collaboration with AtCoder Inc. (AtCoder). We are deeply grateful for their outstanding expertise and contributions in optimization and algorithms, which were invaluable in providing data, analyzing results, and enabling our AI agent’s participation in their contests.

Sakana AI

237,195 Aufrufe • vor 1 Jahr

📢 Join us tomorrow morning at our CVPR 2025 poster session (#340, ExHall D, 10:30am–12:30pm) to chat about Project Magma 👉 This is a big team effort to build a multimodal agentic model capable of understanding and acting in both digital and physical environments—just like how we interact with the world every day. 🚀 Even more exciting: we demonstrate the scaling potential of agent pretraining on large-scale human instructional videos through our Set-of-Mark (SoM) and Trace-of-Mark (ToM), showcasing strong zero-shot performance in: 1) multimodal image/video understanding, 2) UI navigation, 3) Real-world robot manipulation and even 4) Gaming! We've received encouraging feedback over the past few days—and this is only the beginning. A small step forward, with exciting things ahead!

📢 Join us tomorrow morning at our CVPR 2025 poster session (#340, ExHall D, 10:30am–12:30pm) to chat about Project Magma 👉 This is a big team effort to build a multimodal agentic model capable of understanding and acting in both digital and physical environments—just like how we interact with the world every day. 🚀 Even more exciting: we demonstrate the scaling potential of agent pretraining on large-scale human instructional videos through our Set-of-Mark (SoM) and Trace-of-Mark (ToM), showcasing strong zero-shot performance in: 1) multimodal image/video understanding, 2) UI navigation, 3) Real-world robot manipulation and even 4) Gaming! We've received encouraging feedback over the past few days—and this is only the beginning. A small step forward, with exciting things ahead!

Jianwei Yang

13,112 Aufrufe • vor 1 Jahr

🔍 Introducing BackSearch. LLMs are increasingly asked to predict the future, but a good backtest requires a snapshot of the internet at a point in time. BackSearch allows LLMs to search the web as it was on a particular date. It’s great for: 🔮 Forecasting and prediction markets. 📈 Quantitative finance. 🌍 RL environments that simulate the world. 🥶 Freezing websearch for benchmark reproducibility We’re releasing a narrow slice of our index to begin with, focused on the news domain for 2026. Based on feedback we’ll open up more of our index in subsequent releases. 👇 Try BackSearch out with the link below.

🔍 Introducing BackSearch. LLMs are increasingly asked to predict the future, but a good backtest requires a snapshot of the internet at a point in time. BackSearch allows LLMs to search the web as it was on a particular date. It’s great for: 🔮 Forecasting and prediction markets. 📈 Quantitative finance. 🌍 RL environments that simulate the world. 🥶 Freezing websearch for benchmark reproducibility We’re releasing a narrow slice of our index to begin with, focused on the news domain for 2026. Based on feedback we’ll open up more of our index in subsequent releases. 👇 Try BackSearch out with the link below.

General Reasoning

419,098 Aufrufe • vor 1 Monat

Tencent announces AppAgent Multimodal Agents as Smartphone Users paper page: Recent advancements in large language models (LLMs) have led to the creation of intelligent agents capable of performing complex tasks. This paper introduces a novel LLM-based multimodal agent framework designed to operate smartphone applications. Our framework enables the agent to operate smartphone applications through a simplified action space, mimicking human-like interactions such as tapping and swiping. This novel approach bypasses the need for system back-end access, thereby broadening its applicability across diverse apps. Central to our agent's functionality is its innovative learning method. The agent learns to navigate and use new apps either through autonomous exploration or by observing human demonstrations. This process generates a knowledge base that the agent refers to for executing complex tasks across different applications. To demonstrate the practicality of our agent, we conducted extensive testing over 50 tasks in 10 different applications, including social media, email, maps, shopping, and sophisticated image editing tools. The results affirm our agent's proficiency in handling a diverse array of high-level tasks.

Tencent announces AppAgent Multimodal Agents as Smartphone Users paper page: Recent advancements in large language models (LLMs) have led to the creation of intelligent agents capable of performing complex tasks. This paper introduces a novel LLM-based multimodal agent framework designed to operate smartphone applications. Our framework enables the agent to operate smartphone applications through a simplified action space, mimicking human-like interactions such as tapping and swiping. This novel approach bypasses the need for system back-end access, thereby broadening its applicability across diverse apps. Central to our agent's functionality is its innovative learning method. The agent learns to navigate and use new apps either through autonomous exploration or by observing human demonstrations. This process generates a knowledge base that the agent refers to for executing complex tasks across different applications. To demonstrate the practicality of our agent, we conducted extensive testing over 50 tasks in 10 different applications, including social media, email, maps, shopping, and sophisticated image editing tools. The results affirm our agent's proficiency in handling a diverse array of high-level tasks.

AK

343,919 Aufrufe • vor 2 Jahren