ChainCatcher news, the decentralized AI training platform FLock.io has officially reached a strategic partnership with hardware wallet provider OneKey to jointly launch an AI model training challenge focused on smart contract security. This collaboration will integrate FLock's federated learning architecture with OneKey's expertise in blockchain security to jointly promote the upgrade of industry security standards.This challenge is built on the training framework of FLock's core module AI Arena, utilizing an audited real vulnerability dataset, along with a dynamic attack and defense scenario simulation question bank. Participants can engage in the competition by optimizing detection algorithms, contributing threat intelligence, validating model effectiveness, and more, competing for $FLOCK token rewards and OneKey hardware wallets.In addition, the project team has specially opened the "Elite Trainer Program" registration channel, inviting global security engineers and AI developers to collaboratively create efficient and accurate smart contract security detection models.

ChainCatcher news, according to Chainwire, the on-chain AI model verification and rental market Aura has completed a $5.5 million seed round financing, led by Daxos Capital, Manifold Trading, and Selini Capital, with participation from Hermeneutic Investments.With the completion of this round of financing, Aura is accelerating its technology roadmap and preparing for the next phase of platform development.

ChainCatcher news, sources say that DeepSeek is accelerating the launch of its R2 artificial intelligence model, which was initially planned for release in May, but is now working hard to launch it as soon as possible.

ChainCatcher news, according to Jinshi reports, an employee of OpenAI publicly accused Musk's xAI company of having misleading benchmark results for its latest AI model Grok3. In response, xAI co-founder Igor Babushkin insisted that the company did nothing wrong.xAI's chart shows that the two versions of Grok3—Grok3 Reasoning Beta and Grok3 mini Reasoning—outperformed OpenAI's currently strongest available model o3-mini-high on AIME 2025. However, OpenAI employees quickly pointed out on the X platform that xAI's chart did not include the AIME 2025 score of o3-mini-high under the "cons@64" condition.Babushkin argued on the X platform that OpenAI had previously released similar misleading benchmark charts, even though these charts were used to compare the performance of its own models.

ChainCatcher news, the Pando Fund released a research report titled "Decentralized AI Changed by Deepseek," which points out that the narrative of decentralized AI is continuously reshaping. Web3 AI companies can focus on replicating DeepSeek's success while offering new advantages such as multimodality, user ownership, censorship resistance, and privacy. It is expected that the number of projects on the supply side will continue to grow, while consumer-facing projects will begin to compete with Web2 counterparts by building community-participated networks. In the coming year, AI models developed based on a complete Web3 AI stack will emerge one after another.In addition, companies combining AI and crypto are gradually adjusting their strategies to focus on infrastructure building rather than model development. For example, companies in the GPU market like Akash, Render, IoNet, and Exabits have developed sustainable revenue models. Meanwhile, businesses like Grass and Gradient, which allow users to share network bandwidth, have found their market positioning by providing distributed network services to Web2 clients. In inference tasks, the performance gap between small models and large models is narrowing, which means that Web3 does not need to rely on the massive computing power of traditional AI giants and can utilize these streamlined models for efficient inference operations. As this trend develops, we may see more inference endpoints driven by decentralized computing networks in the future.

ChainCatcher news, according to TechCrunch, just hours after xAI released its latest AI model Grok 3, the X platform raised the monthly fee for its top subscription plan Premium+ to $50 and adjusted the annual fee to $350. This is the second price increase for the plan in two months, having previously been raised from $16 per month to $22 in December last year.It is worth noting that some premium features of Grok 3, such as "deep search" and "reasoning," will be charged separately through the SuperGrok plan of the Grok app, and Premium+ users will only have access to basic features. Currently, there are inconsistencies in the pricing displayed on the X platform, with the actual settlement price being $40 per month or $395 per year.

ChainCatcher news, the Solana ecosystem DePIN protocol io.net has announced the launch of the IO Intelligence platform, which combines Vector Database as a Service with up to five intelligent agents and 25 open-source AI models, including DeepSeek, Meta Llama, and Mistral.

ChainCatcher news, according to Jinshi reports, the artificial intelligence community Hugging Face shows that DeepSeek has just released the open-source multimodal AI model Janus-Pro.Among them, Janus-Pro-7B has outperformed OpenAI's DALL-E 3 and Stable Diffusion in the GenEval and DPG-Bench benchmark tests.

ChainCatcher news, ai16z co-founder Shaw posted on platform X, stating that more powerful models are always a good thing for AI agents.For years, AI labs have been surpassing each other in benchmarking and capabilities. Sometimes Google is ahead, sometimes OpenAI, and sometimes Claude. Today it is DeepSeek. The trend is that the largest and most capital-rich companies in the world are competing for a technology that ultimately tends towards being free, open-source, and running on home computers at no cost.The winners of this race have always been hardware and consumer products. Nvidia is always a winner. Every model is optimized for its hardware. Apple is also always a winner: they have invested in a unified memory architecture that allows high VRAM machines to run the latest models (albeit slowly).Products continue to benefit from the latest models. Cursor and Perplexity are examples that magically get better every few months, but as AI integrates into almost every product, all these products benefit from cheaper and faster AI models.AI agents represent a new application paradigm—the core argument is that applications need to migrate to the social media where users are, and agents are a form of application that can exist entirely on social media without users needing to leave. They are self-promoting and benefit from the network effects of every user interaction.When a new model emerges, integrating it into the agent framework usually requires just a few lines of code. Most model providers follow the same API conventions, adhering to OpenAI, so this work can typically be done in minutes. This allows any agent application to have immediate access to the latest models. Whenever a state-of-the-art model appears, agents become smarter.This week is a huge victory for all of us. For agents, for humanity, and for the AI model teams, who now have the motivation to work harder and do better. I am not worried at all about our position in all of this. We are building the next version of Eliza, which will only get better. Thousands of teams are building on our technology, with over 500 contributors to the core repository, and this number will continue to grow as we evolve. We are creating a template for ambitious founders on how to crowdfund their public product projects, and teams will be rolling out more content in the coming weeks and months to solidify this strategy.

ChainCatcher news, the iOS version of the AI model Grok developed by xAI, a company under Musk, has just officially launched on the App Store in the United States.

ChainCatcher message, the Web3 AI platform Tearline has now integrated RootData's data API, enhancing the reliability and richness of its AI models by utilizing the project details, financing, team, and other information provided by the API.It is reported that Tearline is an intent-driven Web3 AI that helps projects and media drive traffic. It can assist users in creating customized content using natural language commands, from social media posts to press releases. Additionally, it can create bots to easily access trending topics.

ChainCatcher news, according to sources, Microsoft has been working to increase both internal and third-party artificial intelligence models to support its flagship AI product Microsoft 365 Copilot, diversifying the existing OpenAI foundational technology and reducing costs.Microsoft is a major supporter of OpenAI, and this time Microsoft is seeking to reduce its reliance on the AI startup, which is different from the way Microsoft has boasted about its access to OpenAI models in recent years.Sources say that in addition to training its own smaller models (including the latest Phi-4), Microsoft is also working to customize other open models to make 365 Copilot faster and more efficient. It is reported that the aim of this move is to lower the costs of running 365 Copilot and potentially pass the savings on to end customers.

ChainCatcher news, according to IT Home, five major media outlets including the Canadian Press, CBC, Torstar, Postmedia, and the Globe and Mail have filed a lawsuit against OpenAI in the Ontario Superior Court, accusing it of extensively scraping their copyrighted content for training AI models like ChatGPT without permission and compensation. The lawsuit document spans 84 pages, demanding OpenAI to compensate for damages and to permanently prohibit unauthorized use of their content. OpenAI responded that its model training is based on publicly available data, adheres to fair use principles, and provides publishers with content attribution and opt-out options.

ChainCatcher news, a source familiar with the matter and two individuals discussing Amazon's plans, stated that Amazon has developed a new generative artificial intelligence that can handle images and videos in addition to text.The model can help customers search for specific scenes in video archives using simple text prompts, such as a game-winning basketball shot. During the development of the new model, Amazon indicated that it still hopes its internally developed AI can attract its cloud customers, thereby reducing reliance on Anthropic's AI, whose Claude chatbot is currently a popular product on Amazon Web Services.Last week, Amazon announced that it would double its investment in Anthropic to $8 billion.

ChainCatcher news, the open-source AI platform Sentient announced on X the launch of the OML 1.0 fingerprint library feature, which supports fingerprint recognition of AI models to verify the identity of AI model developers.Previous news, according to The Block, the open-source AI platform Sentient announced the completion of a $85 million seed round financing, led by Peter Thiel's Founders Fund, Pantera Capital, and Framework Ventures, with participation from Ethereal Ventures, Robot Ventures, LD Capital, Symbolic Capital, Delphi Ventures, Hack VC, Arrington Capital, HashKey Capital, Canonical Crypto, and Foresight Ventures.Additionally, according to the Web3 asset data platform RootData, Sentient is building community-contributed open-source AI models with the goal of directly competing with OpenAI. Sentient will launch activities for contributors, each with specific metrics to evaluate contributions and rewards based on these metrics.

ChainCatcher news, social media and tech company Meta has opened its artificial intelligence model Llama to the U.S. military and defense contractors for national security purposes. Meta's President of Global Affairs Nick Clegg wrote in a statement on November 4 that Llama will be used to streamline complex logistics and planning, track terrorist financing, and strengthen U.S. cyber defenses.The company will collaborate with Microsoft, Amazon, IBM, Oracle, Palantir, and other tech giants to provide comprehensive services to the U.S. government.

ChainCatcher news, FLock.io announced the signing of a memorandum of understanding with Animoca Brands. Both parties plan to jointly develop AI models specifically designed for Animoca Brands, focusing on due diligence, investment, market making, and operational support. The FLock.io platform combines federated learning and blockchain technology, aiming to ensure data privacy through decentralized collaboration and promote collaborative training of AI models.FLock.io CFO Vincent stated that this collaboration will leverage FLock.io's privacy-preserving AI training infrastructure platform to develop four AI models in specific areas for Animoca Brands, while ensuring the localization of core data.