AI News

How do you determine the most powerful AI models currently available? Just take a look at the leaderboards. Community-compiled leaderboards for AI models have surged in popularity online in recent months, providing a real-time window into the jockeying of major tech giants in the AI space. The various leaderboards document which...

High-quality AI inference model goes universal. Early this morning, OpenAI released its new inference model o3-mini. OpenAI calls it its most cost-effective inference model, with significantly improved complex inference and dialog capabilities, outperforming its predecessor in science, math, programming, and other areas o...

I. BACKGROUND AND CHALLENGES With the rapid development of AI technology, large-scale language models (LLMs) have become a core driver in the field of natural language processing. However, training these models requires huge computational resources and time costs, which has led to the rise of Knowledge Distillation (KD) techniques. Knowledge distillation works by ...

Four 10s! It's a rare occurrence, but in ICLR, where the average score is only 4.76, how can it not be considered as quite a blowout? The paper that has won over the reviewers is IC-Light, a new work by ControlNet author Lumin Zhang. we...

What is Refly? Refly is an AI-native content creation platform built on the concept of "free canvas", which provides users with an AI document editor that is visible and accessible through multi-threaded conversations, knowledgebase integration, contextual memories, intelligent search, and...

The new version of Copilot is, seriously, pretty hard to use. It's fancy and doesn't have a point. The results of the query is also no relevant reference page. And "+" sign here to upload only upload images, which for many who want to use copilot to ask questions on PDF files, it is...

Pika recently launched an exciting new feature, PIKAFFECT, which will revolutionize the way users create special effects videos. The feature allows users to apply a variety of unique visual effects to video content, including: Crumble: makes objects look like...

In the field of RAG, the multi-vector model ColBERT brings improvement in retrieval accuracy by generating independent vectors for each token of a document. However, it also brings a dramatic increase in storage requirements and only supports English, which limits its application scope. To solve these problems, we ...

Learn how Rexera migrated to LangGraph to create powerful quality control intelligences for real estate business processes and significantly improve the accuracy of their Large Language Model (LLM) responses. Rexera is revolutionizing manual processes by leveraging AI to automate...

Google recently announced that its video generation model, Veo 2, has been officially integrated into the Gemini Advanced service as well as its experimental platform, Whisk. The move means that Google One AI Premium subscribers can now...

Opera, the Norwegian tech company known for its innovative spirit, is once again at the forefront of browser technology. Recently, Opera announced the launch of a new AI agent called Browser Operator, which brings an unprecedented browsing experience to users. This...

As Chinese AI newcomer DeepSeek makes waves in the global AI space with its open-source, low-cost models, OpenAI CEO Sam Altman is on a low-key trip to Tokyo. At the center of the visit, no doubt, is the industry leader's...

We have released vdr-2b-multi-v1, the best multilingual embedding model for visual document retrieval. We also released its English-only version, vdr-2b-v1, and open-sourced the new vdr-multilingual-train number...

As part of the official launch of Devin, we're offering a Devin Team Plan quota of 500 free ACUs to selected open source project maintainers. Visit app.devin.ai today to sign up and register via osi@cogniti...

January 30, 2025 - Microsoft says DeepSeek's R1 model is now available on its Azure cloud computing platform and GitHub tools for developers in general. Microsoft also said that customers will soon be able to use DeepSeek's R1 models on their Copi...

OpenAI confirmed that the name David Mayer was flagged by an internal privacy tool and said that "in some cases, ChatGPT may not provide some of the information about people to protect their privacy". This past weekend ChatGPT users went through...

Thank you for your patience! We are pleased to announce that the desktop version of ComfyUI, previously known as V1, is now available to beta users! Download ComfyUI Desktop - Windows Download ComfyUI Desktop - Mac About the setup...

On January 31, 2024, OpenAI hosted an AMA (Ask Me Anything, an online question and answer) event on Reddit with a number of OpenAI's core figures, including Sam Altman (CE...

On March 16th, Baidu officially released two new big models: Wenshin Big Model 4.5 and Wenshin Big Model X1, which are already online on Wenshin Yiyan website and users can experience them for free. At the same time, Wenshin Big Model 4.5 is now available on Baidu Intelligent Cloud Qianfan Big Model Platform...

General Introduction Common Sense Education offers a series of free AI literacy courses designed for students in grades 6-12. These lessons are designed to help students understand the basic concepts of Artificial Intelligence (AI) and its social and ethical implications. The curriculum covers AI...

Today, we're expanding Bard's latest technologies - including Gemini Pro in Bard - to more languages and regions. In addition, we've added new image generation capabilities to breathe even more life into your creativity. Available in all supported languages and locales to enable...

SYSTEM DESIGN PRINCIPLES The optimization goals of the DeepSeek-V3/R1 reasoning service are: higher throughput and lower latency. To optimize these two goals, DeepSeek employs a solution called cross-node expert parallelism (EP). First, EP significantly scales the batch ...

🏠 Upgraded Framework Positioning: Bottom-Level Architecture + High-Level Tools - Bottom-Level Advantage: LangGraph has always been characterized by "low-layer, no hidden logic", which is suitable for production environments. Enterprise users (e.g. Uber, LinkedIn) use it to flexibly build customized AI Agen...

Experts generally believe that 2024 is the year of AGI. In this year, the big model industry has undergone radical changes: OpenAI's GPT-4 is no longer out of reach; image and video generation models are becoming more and more realistic; multimodal big language models, reasoning models, and intellectual...

DeepSeek has been hit by a large-scale malicious attack that temporarily restricts new registrations due to an attack on its online service that has resulted in a busy registration process. The issue started to erupt around January 27, 2025 by a deepseek api error report, during which registration also experienced small-scale issues...

How to use the o1 properly: don't write prompts; write briefs, focus on the goal: describe what you want, not how you want to get it, and be aware of the o1's pros and cons! Since the release of o1 in October and o1 pro/o3 in December...

In the era of rapid evolution of artificial intelligence technology, OpenAI will soon launch the highly anticipated "Operator" AI agent tool, which is expected to be released in January 2025 as a research preview. This innovative product marks a new phase of AI from passive response to active execution, and is expected to...

Inception Labs introduces the Mercury family of diffuse Large Language Models (dLLMs) that are up to 10x faster and cheaper than existing LLMs, pushing language modeling to new frontiers of intelligence and speed. Core Essentials Inception...

Open source is on the horizon: the massification of AI video creation Exciting news has arrived! AliCloud has officially announced that its much-anticipated next-generation AI video generation model, WanX 2.1, will soon be open sourced! 🎉 This model, which has been hailed as "redefining video generation", has been released since...

Anthropic's much-anticipated Claude 3.7 Sonnet model was accidentally leaked on the eve of its planned release date, dropping a heavy bomb in the calm AI waters and instantly igniting a wave of attention in the tech industry. According to the current leak...

Many of you have probably heard the jokes about robots taking over the world. These jokes were once based on a seemingly unattainable reality, but today there is real anxiety lurking behind them. Artificial intelligence (AI) is no longer a science fiction concept, but a real and increasingly powerful technology. While the likes of C...

Data from the World Meteorological Organization (WMO) shows that in the last 50 years, an average of one weather, climate, or flood-related disaster has occurred every day, resulting in 115 deaths and $202 million in economic losses every day. Therefore, building more accurate weather forecasting systems...

Where did the PR Assistant come from? PR Assistant is a new player in the software development arena. It came about because people needed to make it easier to manage code changes. It came about because of the need to make it easier to manage code changes, which used to be a pain in the ass for people to do. PR Assistant started as a way to automate simple checks, but has become very...

Luma Labs is once again setting the bar in AI-powered video generation with Ray2, an innovative model based on Luma's groundbreaking multimodal architecture that brings a quantum leap in digital content creation. This innovative model, built on Luma's groundbreaking multimodal architecture, brings a quantum leap in digital content creation, putting realistic, vivid video content at your fingertips.Ra...

Clio: A Real-World AI Usage Insight System for Privacy What do people use AI models for? Despite the rapidly growing popularity of big language models, until now we've lacked insight into exactly how they're used. It's not just a matter of curiosity...

With only 14 billion (14B) parameters, Phi-4 demonstrates performance comparable to or even surpassing some larger-scale models through innovative training methods and high-quality data. In this paper, we describe in detail the architecture, features, and training methods of Phi-4, as well as its practical application in ...

NLR.AI is an innovative platform that showcases the power of autonomous AI creation. The site features a project called Terminal Velocity, a novel created autonomously by 10 specialized AI agents. Each AI agent has a specific role...

According to insider sources, Anthropic is actively working on two new features called Harmony and Compass that are designed to significantly enhance the capabilities of its AI model Claude. These new features are expected to be integrated into Claude, providing users with...

Z Highlights Intuition works in about half of the jobs. For example, when there's a clear product direction and you're just doing the final fine-tuning, trying to understand the target users and the exact problem to be solved, intuition is helpful because the situation is closer to the traditional product launch process. But in the case of item...

If you have used Conch APP, you may have used his voice cloning, the effect is very good, but you can only use it as a voice assistant in the APP. This time, Conch AI video international version official website, new "Audio" channel, you can easily clone your own voice. Back to the main interface of text-to-speech...

AI technology is penetrating into all industries at an amazing speed, and the field of film and television production is no exception. In this wave of integration of technology and art, the rise of AI has brought unprecedented opportunities for change in the film and television industry. Especially in the field of short drama production, which has long faced high production costs, long production cycles...

With the rapid development of artificial intelligence technology, the reasoning ability of large language models on difficult scientific topics at the graduate level has become a hot topic of research. Taking OpenAI as an example, its new model OpenAI o1, officially released in early December, demonstrates strong scientific reasoning ability...

AI Highlights Google's MobileDiffusion generates high-definition images directly on your phone in half a second Microsoft's Face Check uses selfies to enable user identity verification for secure password resets OpenAI for DALL-E 3 ...

Google recently announced an experimental feature for its Gemini AI assistant called Gemini with personalization. The new feature will allow Gemini to connect with...

Guest contributors Lennart Heim and Sihao Huang, this article is cross-posted on Lennart's personal blog, Lennart is a regular contributor to ChinaTalk, and recently participated in a discussion on geopolitics in the era of test-time computing...

Abstract Four artificial intelligence systems - ELIZA, GPT-4o, LLaMa-3.1-405B, and GPT-4.5 - were evaluated by independent populations in two recent randomized controlled Turing tests. The study was conducted by UC San Diego's Cameron R...

Yesterday, Sam Altman, CEO and co-founder of OpenAI, posted his latest in-depth article, Reflections, on his personal blog. The main review of OpenAI's founding in the past 9 years: from the initial unappreciated by the outside world, to the 20...

After 7 months of development, 1 month of testing and 77,376 lines of code, Refly is officially open source! ⚡️🔥🚀 Since the project's inception, Refly has been striving to become a world-class open source project on par with Docker and K8S. Our mission...

GitHub has announced a major new product called GitHub Spark - allowing users to build and share micro-applications using only natural language descriptions, dramatically lowering the barriers to software development. Users don't need to have deep programming knowledge; the system automatically handles all the technical details...

The copilot pages feature is very powerful and is the highlight of Microsoft's launch this year: until now, it was only available to users with a copilot for Microsoft 365 subscription. Today, the copilot pages feature is officially...

Claude.ai has recently launched a new analytics tool designed to enhance its data processing and analytics capabilities. The introduction of this tool marks Claude's transformation as a data analytics assistant, enabling it to write and run JavaScript code that enables complex...

This is Perplexity's second acquisition following the 2023 acquisition of Spellwise, whose CEO was responsible for developing Perplexity's mobile apps. Perplexity's current...

Overview In the age of the information explosion, organizations have come to rely on search technology not just to find content, but to improve efficiency and productivity. However, traditional search models often struggle to truly understand user intent, resulting in inaccurate, irrelevant or even incomplete search results. This experience not only frustrates users...

GitHub has announced a free program for its AI programming assistant, GitHub Copilot, now available to all users in Visual Studio Code. All users need is a GitHub account to start using...

Remember in 2007, Steve Jobs took the first generation of iPhone out of the sky and opened a new era of smartphones? A flash of more than a decade has passed, although the smartphone is becoming more and more powerful, but it seems to have reached the bottleneck of innovation. Just when everyone is lamenting "technology is based on shell change", Op...

Artificial Intelligence (AI) technologies are penetrating all aspects of software development at an unprecedented rate, and the evolution of AI programming tools is particularly compelling. In this paper, we will analyze the trends of the new generation of AI programming tools, and combine them with AutoDev Sketch's practical...

It's finally here! ~ Copilot in Excel with Python Advanced Analysis Python Advanced Data Analysis in Excel is finally available. Been waiting for this button for a week now, and last week Microsoft released this new...

Over the past few years, there has been a never-ending discussion about the next phase of AI development: Agentization? Stronger reasoning capabilities? Or true multimodal fusion? The speculations are endless, signaling a new transformation in AI. Now, I think it's time to give a clear judgment...

The development of smart phones today, the competition for hardware, building application ecology seems to have become the "old script". Now, the new growth point of the cell phone industry, everyone is aiming at the same direction - artificial intelligence. This time, the most popular technology focus, fell on the so-called "Agent (intelligent body)" before...

SmolVLM is a small multimodal model with 2 billion parameters that accepts any combination of image and text input and generates text output. After launching the SmolLM lightweight language model in July, AI app development platform Hugging Face ...

Now that OpenAI has made a major update to the ChatGPT API, we've summarized 8 important points. It is worth mentioning that there is a new "GPT-4-0125-preview" model, which allows for better handling of tasks. 1...

Sonauto has taken an important step in the convergence of artificial intelligence and music technology by officially launching its AI Music Generation API for developers. This innovative API is designed to empower developers around the world to seamlessly integrate powerful AI music creation capabilities into a variety of applications...

Yesterday, DeepSeek released a preview of DeepSeek-R1-Lite, an autonomous reasoning large language model that can compete with o1 and show users the full thought process that o1 does not make public. Similar to OpenAI's o1-pr...

A spokesperson for OpenAI has issued an official statement in response to the incident. The statement said that the Sora model is still in the "research preview" phase and that the company is trying to balance creativity with strong security measures so that the technology can be used more widely in the future. ope...

Mixture of Experts and Sparse attention make virtually unlimited contexts possible. It allows the RAG AI Agent to devour entire codebases without contextual constraints...

Recently, the field of large-scale language modeling has been receiving increasing industry attention for a new paradigm of reinforcement learning in the late stages of training. Following the introduction of O-series models such as GPT-4o by OpenAI and the release of DeepSeek-R1, the outstanding performance of the models proves that reinforcement learning in the optimization process...

ConchQ released v2.9.1 for mobile, adding many exciting features. It's the Chinese New Year, Conch wishes you a Happy Chinese New Year and a prosperous Year of the Dragon! ConchQ is a great gift for Chinese New Year. Upgrade now to unlock new ways to play in Chinese New Year. Clone your voice in 10 seconds to unlock the sound of your heart...

From our Inferentia and Trainium chips built for inference and training, to providing the industry's best Foundation Models at Amazon Bedrock, and AI-driven...

AI headlines have been covered by DeepSeek for more than ten days, and yesterday, OpenAI finally sat down and launched a new inference model series, o3-mini. o3-mini not only opens up inference models to free users for the first time, but also reduces the cost of inference models compared to the previous o1-series, which was...

Today's AI Revealed: daily top and featured AI advances. The Grammys take on the AI challenge, too. The world's first AI child is introduced. AI is quietly changing our daily lives. Daily AI Tool Picks Zipchat AI is the only 100 percent...

At Inflection, our mission is to create a personal AI for everyone.Last May, we launched [Pi] - a personal AI that is empathetic, helpful, and safe.In November, we announced the launch of a new primary base model [Inflectio...

Have you ever struggled to configure AI Assistant features? In the past, adding new features to AI Assistant often required tedious steps such as typing complex commands into the terminal, manually resolving dependencies, consulting technical documentation, and praying that everything went smoothly. This complexity has left many users eager to ...

This site does not provide download address and use tutorial, please solve it by yourself.W A R P + security unknown, mainly used for git project download and login web version ChatGPT. Full use tutorial: https://zhuanlan.zhihu.com/p...

Anthropic has announced that its Claude artificial intelligence (AI) chatbot can now respond in a variety of styles to better match a user's communication preferences, tone and structure. New styles to choose from include formal, concise, and explanatory, in addition to customizable styles. If...

OPENAI has just released its own AI search product, SearchGPT, which is still in the prototype stage and needs to be applied for a waiting list before it can be used after waiting for review. AI search is a feature that OPENAI promised to launch in April, and it wasn't until July that the prototype was launched, a full delay...

Data analysis is a critical but challenging task for business professionals. From creating and preparing data to gaining insights and performing advanced analysis, the process can be both time-consuming and complex. Introducing Copilot in Excel: a powerful work designed to simplify and transform data analysis...

Frameworks like LangChain, CrewAI, and AutoGen have become popular by providing high-level abstractions for building AI systems. However, many developers, including myself, have found that these tools do more harm than good, often adding unnecessary complexity and frustration to the development process...

Conclusion Domestic primary and secondary schools have issued documents to popularize AI education from top to bottom, and the mature stage of the "industry" is to get certificates, advancement, training, and finally become a rich man's game. It may be better to follow the example of the United States and directly enter the experimental stage of popularization of science or learn from Japan to give a clear guiding learning framework for the early stage of practice...

Artificial intelligence-generated content is growing at an unprecedented rate, with four of the 20 most popular posts on Facebook last fall reportedly generated by AI. Additionally, Medium estimates that 47% of content on its platform also comes from AI. as...

With the rapid development of artificial intelligence technology, large-scale language models (LLMs) are changing our lives at an unprecedented rate. However, technological advances also bring new challenges: LLMs can be maliciously exploited to leak harmful information or even be used to create chemical, biological, radiological, and nuclear weapons...

This document is a PPT of a talk given at Stanford University by Barret Zoph and John Schulman, OpenAI's pre- and post-training leaders (and OpenAI co-founders), sharing their experience in OpenAI's development of Ch...

Highly regarded AI giant OpenAI recently announced the release of a research preview of its latest flagship language model, GPT-4.5, once again generating a lot of attention from the tech community. The highly anticipated new model will initially be available to software developers and ChatGPT Pro...

Mistral Small 3: Apache 2.0 Protocol, 81% MMLU, 150 tokens/sec Today, Mistral AI launched Mistral Small 3, a latency-optimized 24 billion...

Google has released what it's calling a new "reasoning" AI model - but it's still in the experimental stage, and from the looks of our brief tests, it does have room for improvement. The new model is called Gemini 2.0 Flash Thinking E...

The smallest model in our R family delivers top-notch speed, efficiency, and quality to build powerful AI applications on common GPUs and edge devices. Today, we are excited to release Command R7B, our large language model (LLM) developed specifically for enterprise...

Recently, the artificial intelligence research organization OpenAI quietly launched a new online education platform called OpenAI Academy without large-scale publicity. The platform is designed to provide free AI-related learning resources to users around the world, marking the launch of O...

Introduction: What is Conch Ask? Hello everyone, let me introduce you to this amazing contraption - Conch Ask AI.Conch Ask, it's not a seafood product hawked by any seaside vendor, but an AI intelligent chatting assistant from Shanghai Ruyu Technology Co. Don't underestimate this guy, it...

In late 2024, YC Partner Jared predicts that vertical AI Agents will be an emerging market 10x larger than SaaS in the next few years, and that the space could also spawn tech giants with market caps of over $300 billion. By then, Microsoft CEO Saty...

ChatGPT can now search the web better than ever before. You can get fast, timely answers with links to relevant web resources that you previously had to search through a search engine to find. This combines the benefits of a natural language interface with the latest sports scores, news, stock quotes...

Today, DeepSeek, a rising star in China's AI field, has triggered an "earthquake" in the global science and technology industry with its amazing speed and strength. This application, which is known as "the light of domestic AI", not only pushed ChatG...

With the Wenxin intelligent body function tends to stabilize, this time to turn out a on-line is not long, but the network has been little introduction of the Baidu marketing tool "clever cabin". Do Baidu bidding friends should be recommended by the clever cabin marketing tool, especially industry customers, some people treat him as intelligent customer service, some people only as optimization r...

✨ Little Red Book Nuggets Secrets Revealed! 🔥 Leek programs can be fun? There are also operation tutorials for free! Hello everyone, recently found that everyone is very interested in the small red book related projects, it seems that everyone wants to dig gold in the small red book this blue sea ah! 🚀 Don't worry, today I bring you...

Google releases new models in the Gemini 2.0 Flash series, including three new models, Gemini 2.0 Flash, Flash-Lite and Pro, designed to provide developers with faster, more affordable and more powerful generative...

Recently, artificial intelligence company Anthropic has been in the spotlight. Not only has it launched the powerful Claude 3.7 Sonnet model, but it has also made significant progress on the funding front.Mike Krieger, Anthropic's chief product officer...

If you own a GeForce RTX 30-series or newer GPU, a demo app called Chat with RTX enables personalized content services that don't require an Internet connection. Nvidia has launched Chat with RTX, an a...

Today we will delve into the world of open source language models (LLMs) and learn about a new contender for this innovation - Code Llama 70B using OctoAI. Developers often find it difficult to efficiently incorporate a code like Meta's Code Llama 70B...

Here's a story that was previously missed: this week, Nvidia's market capitalization surpassed that of Amazon and Google's parent company, Alphabet, jumping up to become the world's third most valuable company, with a staggering $1.83 trillion market cap. Here's a fun fact: Nvidia last surpassed Amazon in market capitalization...

General Introduction Text2Edit is an open source project , hosted on GitHub , aims to provide efficient text editing and advertisement generation functions . The main goal of the project is to help users quickly process text content and generate high-quality ads through an easy-to-use interface and powerful ...

Would you like to have a podcast auto-played directly from your phone's news feed based on your interests? Google is testing a new feature called Daily Listen...