On February 20th, Elon Musk announced an important update on the social platform X, stating that the Grok application has successfully launched the early testing version of Grok 3 voice mode. This news immediately attracted widespread attention.

Baidu's Wenxin Intelligent Agent Platform has announced full integration with the DeepSeek model, now available for free to all users. The Wenxin Intelligent Agent Platform, launched by Baidu, is a platform for building intelligent agents based on the Wenxin large model, aimed at helping users quickly create and deploy intelligent agent applications.

Ovis2 is the latest version of the Ovis series models proposed by Alibaba's international team. Compared to the previous version 1.6, Ovis2 has significant improvements in data construction and training methods. It not only enhances the capacity density of small models but also greatly improves chain of thought (CoT) reasoning capabilities through instruction fine-tuning and preference learning. Additionally, Ovis2 introduces video and multi-image processing capabilities, and enhances multilingual abilities and OCR capabilities in complex scenarios, significantly increasing the model's practicality.

The AI video field welcomes a game-changing new product! Recently, Google's latest AI video model, Veo 2, was officially released and chose the well-known material platform Freepik for its global debut. This launch not only marks another significant breakthrough in Google's AI video technology but also comes with the exciting news that the first 10,000 users can experience this 'black technology' for free, undoubtedly delivering a substantial benefit to video creators! According to official tweets, Veo 2 is hailed as Google's...

The next-generation AI application building platform, Coze, has officially announced its exclusive support for the DeepSeek Function Calling tool capabilities. The core highlight of this update is the deep integration of the DeepSeek model. Users can not only experience the R1 and V3 models of DeepSeek for free, but also view the thought chain content of the models in real-time through the Coze platform. This feature allows developers to intuitively observe the functionality in scenarios such as the debugging area and the Coze store.

Recently, Figure founder and CEO Brett Adcock announced a new machine learning model called Helix, aimed at enhancing the capabilities of humanoid robots in household environments. This news comes just two weeks after Figure announced the end of its collaboration with OpenAI, demonstrating their strong commitment to the field of robotics technology. Helix is a 'general-purpose' visual-language-action (VLA) model capable of processing visual data.

According to foreign media reports, informed sources revealed that Microsoft's engineering team is accelerating the deployment of server resources to prepare for OpenAI's upcoming GPT-4.5 and GPT-5 models. Although OpenAI CEO Sam Altman recently confirmed that GPT-4.5 will be unveiled in a few weeks, Microsoft internally expects to complete the hosting deployment of the new model as soon as next week. This move indicates that the technological race in the field of generative AI has entered the sprinting stage. Code-named Orion, GPT-4

In the AI industry, Together AI recently announced the completion of a $305 million Series B funding round, which has garnered widespread attention. The company's rise is closely related to its newly launched deep reasoning model, DeepSeek-R1. Contrary to initial concerns, many industry experts believe that advancements in deep reasoning have not diminished the demand for infrastructure; rather, they are continuously enhancing this demand. Image source note: Image generated by AI, image licensed from service provider Midjourney since its establishment in 2023.

In the field of robotics, the American startup Figure AI recently released a large edge model named Helix, marking a significant breakthrough in humanoid robot control technology. Helix is the first visual-language-action (VLA) model capable of high-frequency, continuous control over the entire upper body of a humanoid robot, including the head, torso, wrists, and fingers. This means that the robot can perform tasks directly based on natural language instructions without extensive training.

DingTalk AI Corporate Search, launched in 2024, focuses on addressing the pain points of enterprise users related to information fragmentation and low retrieval efficiency. This tool leverages the understanding, reasoning, and generating capabilities of large models to integrate unstructured data such as chat logs, documents, knowledge bases, schedules, and logs into a structured knowledge network, improving search efficiency by 300% over traditional methods. By offering this free access, DingTalk aims to further assist enterprise users in building dynamic knowledge networks and optimizing workflows.

Recently, the successful launch of the VLM-R1 project has brought new hope to this field. This project represents the successful migration of the R1 Method from the DeepSeek team into visual language models, indicating that AI's understanding of visual content will enter a whole new phase. The inspiration for VLM-R1 comes from last year's open-source R1 Method by DeepSeek, which leverages GRPO (Generative Reward Processing Optimization) reinforcement.

On February 20, 2024, Tencent's AI Smart Workbench product 'ima.copilot' (abbreviated as 'ima') officially launched its mobile application, first on the Android platform, with the iOS version set to be released soon. To enhance user experience, Tencent has also expanded cloud storage space for free to 2GB, further enriching its multi-terminal ecosystem. Since its release in October 2024, the ima product, with its knowledge base-centered design philosophy, has covered Windows and Mac.