Search AI Products and News

Black Forest Labs & Krea launch open-source image model FLUX.1-Krea, addressing AI's artificial look via 12B-parameter diffusion. Enhanced training improves realism in lighting/colors, compatible with FLUX ecosystem for commercial use.....

San Francisco AI startup Deep Cogito launches the Cogito v2 series of open-source large models, focusing on machine intuition and self-improvement capabilities. The series includes four dense and mixture-of-experts models with 70B to 671B parameters. The flagship 671B MoE model has a reasoning efficiency that is 60% higher than DeepSeek R1, and its performance is close to Claude4Opus. Its core technology involves internalizing reasoning paths into model weights through introspective reasoning, enabling continuous evolution while in use. Tests show that the model can

Google released the AI model AlphaEarth Foundations, revolutionizing Earth observation through virtual satellite technology. The model integrates multi-source data such as satellites and radar daily, dividing the Earth's surface into 10-meter grids for long-term tracking, and uses color coding to intuitively present properties like vegetation and land surfaces. Its innovative compression technology reduces storage requirements to 1/16, significantly lowering costs. It has been applied in areas such as crop monitoring and deforestation tracking, with excellent test performance. The dataset is currently available on Google Earth Engine

The StepZen team announced that their latest generation foundational large model, Step3, is officially open-sourced. Step3 is a model designed for enterprises and developers who pursue the optimal balance between performance and cost, aiming to create the most suitable model for the inference era. The open-source address of this model includes GitHub, Hugging Face, and ModelScope, allowing developers to freely download and experience it. Step3 adopts an MoE architecture, with a total parameter count of 321 billion and an activated parameter count of 38 billion. It

Tencent's X-Omni AI model integrates reinforcement learning to unify image generation and understanding, outperforming rivals in text rendering and visual tasks without classifier guidance.....

Quest Mobile reports top 20 fastest-growing apps (June 2025). AIGC's DeepSeek leads with 163M MAU, Doubao hits 141M (+410% YoY). Hongguo (212M) and UnionPay (188M) lead short videos & payments. List reflects AIGC and online video growth potential.....

Chinese robot manufacturer Unitree has launched the world's first full-size humanoid robot R1 priced below $6,000, at only 39,900 yuan. This 25-kilogram robot features 26 joints and can perform high-difficulty movements such as backflips, equipped with a multi-modal AI system. Compared to similar products, the R1 is only one-third the price of Tesla's Optimus, thanks to China's supply chain advantages and lightweight design. Unitree's move comes as it prepares for an IPO, potentially triggering a price war in the industry. Although the R1 is not yet suitable for home use,

Shanghai AI Lab releases the open-source scientific multimodal large model Intern-S1, surpassing existing top models in scientific research fields such as chemistry and materials. The model has cross-modal scientific analysis capabilities, able to interpret complex data such as molecular formulas and protein structures, and also launches a dedicated scientific discovery platform called Intern-Discovery. Intern-S1 will promote the development of multi-agent disease research systems and continue to open-source to support scientific progress.

Ant Group announced the financial reasoning large model Agentar-Fin-R1 at the World Conference on Artificial Intelligence. Built upon Alibaba's Tongyi Qianwen, the model focuses on the professionalism and security of the financial industry. It has shown excellent performance in evaluations such as FinEval1.0, offering 32B and 8B parameter versions. CEO Zhao Wenbiao pointed out that financial large models are key to the integration of AI and finance, and a corpus covering six major categories and 66 subcategories of financial scenarios has been built. The model is currently applied in a bank's AI mobile banking app, leading to a 25% increase in monthly active users.

Step Star releases the next generation Step 3 large model, which will be open-sourced on July 31st. This 321B parameter multi-modal reasoning model performs outstandingly in intelligence and reasoning efficiency, supporting cross-domain knowledge understanding and visual analysis. The company has established the Model Chip Ecosystem Innovation Alliance in cooperation with chip manufacturers and reached a strategic cooperation with Shanghai State Capital, promoting the application of large models. Step 3 achieves 300% reasoning efficiency of competing products on domestic chips, with over 70% improvement in throughput, and has established a 1+N model matrix covering multiple fields. The model has achieved SOTA on multiple rankings.

Step Zenith has officially launched its new research assistant, 'Step Deep Research,' which is now available for experience on the Step Zenith App and stepfun.com platform, while also opening internal invitation testing. This intelligent tool is designed for fields such as finance, consulting, healthcare, law, policy, and academic research. It can independently complete complex research tasks in about ten minutes and generate reports that are both in-depth and professional, providing efficient support to researchers. The core advantage of 'Step Deep Research' lies in its powerful information processing and integration capabilities. In the context of Sequoia China

The SWE-bench team has launched a lightweight open-source project called mini-SWE-agent, which can fix 65% of GitHub project bugs with just 100 lines of code. This version simplifies the architecture, removes complex tool interfaces, and executes commands through Bash, offering better compatibility and fewer dependencies. While retaining the core functions of the original version, it provides tools such as batch inference and a trajectory browser, supporting command-line startup and visual monitoring. The team states that mini-SWE-agent is suitable for developers who prefer simplicity, while the original version is more appropriate for others.

Ali International AI Solution Marco wins the SAIL Star Award at the 2025 World Artificial Intelligence Conference. The solution supports more than 40 languages and covers over 60 e-commerce scenarios, with a hallucination rate of only 0.02%, significantly better than industry standards. Ali International AI service has a daily call volume exceeding 1 billion, and has been opened to external ecosystems, driving a 23-fold increase in AI call volume for partners. It has now launched three Agent applications: Smart Refund, Customs Code, and Online Recruitment, achieving a 15% reduction in refund costs, a 23% improvement in accuracy, and improved recruitment efficiency.

VLM2Vec-V2, a multimodal embedding framework developed by Salesforce Research and other institutions, breaks through the limitations of traditional models. Based on the Qwen2-VL architecture, this framework innovatively unifies image, video, and visual document retrieval tasks. It adds five new evaluation tasks to expand the MMEB dataset. With key technologies such as dynamic resolution and M-RoPE, it achieved an average score of 58.0 across 78 datasets, leading in performance, especially in video tasks. Although its document retrieval is slightly behind ColPali, it

China has taken a leading position in the global artificial intelligence field, contributing 1,509 large models (40% of the global total of 3,755). The popularity of generative AI applications has reached 80.9%, and the industry scale is expected to exceed 700 billion yuan by 2024, maintaining a growth rate of over 20%. Domestic AI products are deeply applied in areas such as office, education, and industry, and are accelerating overseas expansion, such as DeepSeek, which reached 30 million users within 20 days of its launch. China accounts for 60% of global AI patents, demonstrating strong innovation capabilities, with a promising future development outlook.

China leads global AI with 40% of models. GenAI breakthroughs expected by 2025, with 80.9% adoption. Market to hit $100B+ in 2024, growing 20%+. DeepSeek hit 30M DAU, China holds 60% of AI patents, expanding globally.....

Shengshu Technology launched the Vidu Q1 Refiner video feature at WAIC 2025, revolutionizing the video production process. This feature allows users to upload reference images along with text prompts to directly generate videos, simplifying the traditional production workflow. It supports up to seven subjects simultaneously while maintaining consistency. It uses the U-ViT architecture combined with diffusion model technology to address commercial subject consistency issues. Meanwhile, the company has collaborated with Tsinghua University to introduce the embodied intelligence model Vidar, enabling low-cost robot action conversion through a video large model. The company is still focusing on improving...

SenseTime to launch Rixin v6.5, embodied AI platform, showcasing multi-modal & agent capabilities. Aims to integrate AI with physical world, boost productivity & interaction. Partners with ecosystem for AI industrialization & introduces 'Compute Mall' platform.....

iFlytek's upgraded Spark X1 model enhances translation, reasoning, and text generation, achieving top-tier global performance with 20% better translation quality. It scores over 80 in EN-CN conference translation, reduces AI hallucinations, and boosts efficiency for 100+ firms by 50%. Available via web, app, and API.....

iFlytek announced the official launch of the upgraded version of its deep reasoning large model, iFlytek Spark X1, which is trained using fully domestically developed computing power. This upgrade represents a comprehensive advancement, maintaining an edge over the latest versions of top domestic and international large models such as OpenAI's o3 in overall performance. It has also made significant progress in areas such as hallucination management, multilingual capabilities, and voice simultaneous translation, providing users with a smarter, more reliable, and more efficient AI assistant. The upgraded version of iFlytek Spark X1 has significantly improved capabilities in translation, reasoning, text generation, and mathematics.