Group 1 - The core viewpoint of the article highlights the recent funding rounds completed by the Chinese AI startup, Mianbi Intelligent, which focuses on edge large model development, indicating strong investor confidence and growth potential in the AI sector [1][2] - Mianbi Intelligent has successfully completed three rounds of financing since 2024, with significant investments from various funds, which will help establish a robust technological and product barrier for their efficient large model technology [1] - The company aims to accelerate industry empowerment and ecological expansion, promoting the large model's application across various sectors [1] Group 2 - In early 2025, the global AI competition has intensified, with Mianbi Intelligent leading the way in developing high-efficiency large models that offer better performance, lower costs, and reduced power consumption [2] - The launch of Mianbi's first edge full-modal model, MiniCPM-o 2.6, showcases its innovative capabilities, including real-time audio-visual processing and natural language generation, positioning it at the forefront of the industry [2] - The MiniCPM series has achieved over 10 million downloads across all platforms, reflecting its popularity and effectiveness in the market [2]

Group 1: Company Overview - The company, Mianbi Intelligent, has completed a new round of financing amounting to several hundred million yuan, with investments from Hongtai Fund, Guozhong Capital, Qingkong Jinxin, and Moutai Fund [1] - Mianbi Intelligent was founded in August 2022 and is incubated by Tsinghua University's NLP laboratory, with co-founder Liu Zhiyuan serving as the chief scientist [1] - Since 2024, Mianbi Intelligent has completed three rounds of financing, with the last round also being several hundred million yuan [1] Group 2: Market Strategy - Mianbi Intelligent adopts a "small to big" strategy, focusing on low-cost, relatively small parameter models to achieve high efficiency, which has gained recognition in the industry [2] - The company is one of the early adopters of the "edge large model" strategy, which refers to AI models that run locally on devices like smartphones and PCs, independent of cloud servers [2] - The edge large model is expected to accelerate penetration into the industry by 2025, with multi-modal interaction scenarios maturing and widespread applications in consumer electronics [3] Group 3: Commercialization Progress - Mianbi Intelligent's edge large model platform, MiniCPM series, has surpassed 10 million downloads [5] - The company launched its first edge full-modal model, MiniCPM-o 2.6, with a parameter scale of only 8 billion, capable of image understanding and multi-modal real-time interaction [5] - Mianbi Intelligent has made significant strides in the automotive sector, with the first mass-produced model featuring its edge model, the Changan Mazda MAZDA EZ-60, hitting the market [6] Group 4: Industry Impact - The edge AI market in China is projected to grow to approximately 1.9 trillion yuan by 2028, driven by explosive demand for intelligent and real-time capabilities in edge devices [4] - The deployment of AI large models on the edge addresses issues related to network latency, privacy, and computing costs, unlocking the computational potential of devices [4] - Mianbi Intelligent has also ventured into vertical fields such as law and education, assisting in the development of AI systems for judicial processes and educational support [6]

Group 1 - Beijing Mianbi Intelligent Technology Co., Ltd. has completed a new round of financing amounting to several hundred million yuan, with investments from Hongtai Fund, Guozhong Capital, Qingkong Jinxin, and Moutai Fund [1] - Since 2024, the company has successfully completed three rounds of financing, which will further strengthen its foundation for high-efficiency large model technology, product barriers, and accelerate industry empowerment and ecological expansion [1] - The global AI competition is intensifying in 2025, with innovation paths characterized by "high efficiency and low consumption" leading the global AI transformation [1] Group 2 - In January 2025, Mianbi Intelligent launched its first end-side full-modal model "MiniCPM-o 2.6," achieving real-time interaction with an 8B scale and introducing features like "continuous watching, real-time listening, and natural speaking" [2] - The MiniCPM series has achieved over 10 million downloads across all platforms, recognized for its high efficiency and low cost, and has been compared to ChatGPT and GPT-4V [2] Group 3 - Mianbi Intelligent is rapidly advancing its commercial layout and business around high-efficiency large models and end-side AI, exemplified by the launch of the "MiniCPM Super Assistant cpmGO," the world's first pure end-side intelligent assistant for vehicles [3] - The first mass-produced model equipped with the end-side model, the Changan Mazda MAZDA EZ-60, made its global debut in April 2025, marking a new phase in the commercialization of end-side large models in the automotive cockpit sector [3] - The company has established deep collaborations with leading automotive manufacturers and tech giants like Qualcomm and Intel to promote the widespread implementation of native intelligent cockpits [3]

Group 1: Financing and Investment - Recently, Mianbi Intelligence successfully completed a new round of financing amounting to several hundred million yuan, led by Hongtai Fund, Guozhong Capital, Qingkong Jinxin, and Moutai Fund [1] - Since 2024, Mianbi Intelligence has maintained a steady financing pace, completing three rounds of financing, with significant investments from various venture capital firms [1] - Mianbi Intelligence and Zhipu AI are among the few companies that have successfully secured continuous financing amidst a challenging environment for large model companies [1] Group 2: Commercialization in Automotive Sector - Mianbi Intelligence is accelerating its commercialization process, particularly in the automotive industry, with the launch of the "Little Steel Cannon Super Assistant cpmGO," the world's first pure edge-side intelligent assistant for vehicles [2] - The debut of the MAZDA EZ-60, equipped with Mianbi's edge-side model, marks a new phase in the commercialization of edge-side large models in automotive cockpits [2] - The company has established partnerships with leading automotive manufacturers such as Changan Automobile, SAIC Volkswagen, and Great Wall Motors to promote large-scale deployment of edge-native intelligent cockpits [2] Group 3: Expansion into Vertical Industries - Mianbi Intelligence is also making strides in vertical sectors such as law and education, contributing to the development of national-level legal AI infrastructure [3] - The company has assisted in the launch of a judicial vertical large model that has supported over 291,000 cases and generated 11,600 draft documents since its trial run in January 2024 [3] - In education, Mianbi has partnered with Tsinghua University to introduce an AI learning assistant, aiming to create an automated classroom model with a student graduation rate exceeding 40% [3] Group 4: Technological Advancements - Mianbi Intelligence's edge-side model capabilities are continuously evolving, with the MiniCPM series gaining widespread recognition [4] - The MiniCPM-o 2.6 model features 8 billion parameters and supports innovative functionalities such as real-time interaction, achieving international leadership in image understanding and speech processing [4] - The MiniCPM series has surpassed 10 million downloads, establishing itself as a technical benchmark for edge-side large models globally [4]

Core Insights - Recently, Mianbi Intelligent completed a new round of financing amounting to several hundred million yuan, led by Hongtai Fund, Guozhong Capital, Qingkong Jinxin, and Moutai Fund, marking the third round of financing since 2024 [1][2] - Mianbi Intelligent has rapidly developed a complete matrix of full-modal, multi-modal, and foundational models, continuously pushing the boundaries of edge large model capabilities [1] - The MiniCPM series has achieved over 10 million downloads, recognized as the most downloaded and popular Chinese large model on Hugging Face in 2024 [1] Financing and Investment - The recent financing will further establish Mianbi Intelligent's efficient large model technology and product barriers, accelerating industry empowerment and ecological expansion [2] - The company aims to promote the large-scale application of "edge brains" across various industries by collaborating with upstream and downstream sectors [2] Product Development - In September 2024, Mianbi Intelligent released the MiniCPM 3.0 model, outperforming GPT-3.5 with 4 billion parameters [1] - The MiniCPM-V 0.6 model, launched in August 2024, achieved state-of-the-art results in single-image, multi-image, and video understanding with only 8 billion parameters, matching GPT-4V capabilities [1] - The first full-modal model, MiniCPM-O 2.6, was introduced in January 2025, enabling real-time interaction with 8 billion parameters [1] Applications and Collaborations - Mianbi Intelligent launched the "MiniCPM Super Assistant cpmGO," the world's first pure edge intelligent assistant for vehicles [2] - The company participated in the development of the "Faxin Legal Foundation Model," which has been released by the Supreme People's Court [2] - In collaboration with Tsinghua University, Mianbi Intelligent introduced the AI Student Growth Assistant "Qingxiaoda," providing personalized intelligent assistants for all undergraduate students [2]

Core Viewpoint - The article discusses the introduction of the EdgeInfinite algorithm by vivo AI Research Institute, designed to efficiently process long texts on edge devices, overcoming computational and memory limitations associated with existing models [1][4]. Summary by Sections EdgeInfinite Algorithm Overview - EdgeInfinite is specifically tailored for edge devices to handle long text inputs, such as call summaries and personal document summaries, which current models struggle with due to resource constraints [4]. - The algorithm integrates a trainable gated memory module into the Transformer architecture, allowing it to maintain compatibility while achieving good performance on long text tasks with minimal parameter tuning [4][10]. Architecture Components - The architecture consists of three core components: 1. **Chunked Attention Module with ROPE**: It segments the input text into smaller chunks, calculating Q, K, and V values for each, while incorporating position encoding to enhance attention accuracy [8]. 2. **Memory Compression and Decompression**: This module stores past KV states as fixed-length memory blocks, enabling the model to compute attention between current Q states and past KV states, approximating the attention calculation of original long sequences [9]. 3. **Adaptive Gated Memory Module**: It combines memory-based attention with local chunk-based attention, enhancing the model's ability to handle long-distance dependencies [10]. Inference Strategies - EdgeInfinite employs two inference strategies: 1. Retaining specific token KV caches (sink tokens and window tokens) to preserve semantic and positional information for high-quality outputs [13]. 2. A routing mechanism for long and short text tasks, allowing dynamic integration with existing base models to enhance long text capabilities without compromising short text performance [13]. Experimental Results - The performance of EdgeInfinite was tested using the BlueLM-3B model on the LongBench dataset, comparing it with three KV cache optimization methods and the original FullKV model [14]. - Results indicate that EdgeInfinite shows significant advantages in multi-document QA and few-shot learning tasks, outperforming other methods in several instances and maintaining competitive overall model performance [15]. Ablation Studies - Ablation studies confirmed the importance of retaining specific tokens during inference, showing that removing sink or window tokens significantly impacts performance [17]. - Compared to the original BlueLM-3B model, EdgeInfinite demonstrates shorter first-word output times and lower memory usage, maintaining stable memory consumption even with increased input lengths [17]. Future Applications - EdgeInfinite is expected to be widely applicable in resource-constrained devices, enhancing the efficiency of various long text processing tasks, such as in smart voice assistants and mobile document processing, providing users with a smoother experience [17].

Core Insights - The smartphone industry is undergoing an AI revolution, with manufacturers increasingly integrating AI features into their new products, marking a shift from traditional hardware innovation to AI-driven functionalities [2][5][14] - IDC forecasts a dramatic increase in AI smartphone shipments in China, with a year-on-year growth of 591% in 2024, and a penetration rate rising from 3% in 2023 to 22% [4] - The competition among smartphone manufacturers is shifting from hardware specifications to AI capabilities, emphasizing the need for end-to-end AI design from chips to operating systems [8][13] Group 1: Industry Trends - The AI smartphone market is expected to reach 1.18 billion units by 2025, accounting for 40.7% of the overall market [4] - High-end smartphones priced above $600 are projected to exceed 30.9% of the market share, with AI features contributing 75% of their premium pricing [4] - The average replacement cycle for smartphones has extended to 51 months, prompting manufacturers to focus on AI to drive consumer upgrades [5] Group 2: Technological Developments - The new generation of smartphones must feature advanced AI capabilities, including large model computing power, system-level AI integration, and proactive service in various scenarios [8][16] - AI's impact on imaging technology is significant, with innovations allowing for real-time analysis and optimization of images, enhancing capabilities beyond traditional photography [10][11] - The relationship between hardware manufacturers and AI developers is evolving, with companies like Qualcomm and Huawei creating ecosystems that support AI development and deployment [17][22] Group 3: Competitive Landscape - Major smartphone manufacturers are divided into three camps: Apple, Huawei, and an open ecosystem represented by brands like Xiaomi and Honor, each pursuing different AI strategies [20][22] - Huawei is positioned to lead in the AI smartphone market due to its strong R&D investment and technological capabilities in AI chipsets and cloud collaboration [22][23] - The future of smartphones may not solely rely on traditional devices, raising questions about the evolution of AI-native smart devices beyond current smartphones [23][24]

Core Insights - The article introduces Mixture-of-Lookup-Experts (MoLE), a new architecture designed to optimize the deployment of Mixture-of-Experts (MoE) models, particularly in resource-constrained environments [1][28] - MoLE addresses the challenges of high memory usage and transmission delays associated with traditional MoE during inference by replacing matrix operations with lookup tables [28] Group 1: MoLE Architecture - MoLE activates only a small subset of experts needed for each token during inference, significantly reducing computational load while maintaining a large parameter scale [1] - The architecture allows for the pre-computation of input-output mappings stored as lookup tables, enabling efficient retrieval during inference [5][6] Group 2: Training Phase Differences - In the training phase, MoLE modifies the input to routed experts from the previous layer's output to shallow embedding tokens, facilitating the pre-computation and storage of lookup tables [8] - MoLE employs an activation strategy that activates all routed experts during training, eliminating the need for sparse activation to control computational load [9] - The loss design in MoLE focuses solely on language modeling loss, without additional load balancing loss terms [10] Group 3: Inference Phase Process - During inference, MoLE constructs lookup tables from the embedding layer's weight matrix, allowing for direct retrieval of expert outputs based on token IDs [15] - The lookup table is stored in lower storage devices, and during inference, the corresponding expert outputs are retrieved and loaded into memory for computation [16] Group 4: Performance and Efficiency - MoLE's computational complexity during inference is comparable to dense models and traditional MoE models, while significantly reducing transmission overhead [17] - Experimental results indicate that MoLE achieves performance on par with MoE while drastically reducing transmission costs by over a thousand times [20][28] Group 5: Experimental Results - The experiments conducted on the Pile dataset show that MoLE maintains performance equivalent to MoE while using the same training parameters and inference activation parameters [20] - MoLE demonstrates lower inference latency compared to MoE, especially in batch decoding scenarios, highlighting its advantages in high-throughput tasks [28]

Core Viewpoint - The article highlights the rapid advancements in automotive AI technology, particularly focusing on the end-side large model developed by Mianbi Intelligent, which has achieved remarkable speed and efficiency in vehicle applications, revolutionizing the industry standards for AI integration in cars [1][14]. Group 1: Product Launch and Features - Mianbi Intelligent's cpmGO, a pure end-side large model-driven intelligent assistant, was showcased at the Shanghai Auto Show, marking a significant milestone in automotive AI [9][12]. - cpmGO boasts features such as 91% execution accuracy, local data processing, and robust performance in weak network conditions, making it a pioneering product in the industry [10][28]. - The product integrates multi-modal perception and interaction, allowing users to control vehicle functions through voice commands with high accuracy [30][31]. Group 2: Technological Innovations - The cpmGO model is powered by the MiniCPM, which operates entirely on the vehicle's local system, ensuring data privacy and rapid response times [27][28]. - The system's GUI Agent can understand and execute screen commands, enhancing user interaction by performing tasks autonomously based on context [33][36]. - The collaboration with major chip manufacturers like Qualcomm and Intel supports the optimization of cpmGO across various platforms, ensuring compatibility and performance [11][13]. Group 3: Industry Impact and Future Trends - The article discusses the shift in the automotive industry towards end-side AI models, which are less dependent on cloud services, addressing issues like latency and data security [38][42]. - The partnership between Mianbi Intelligent and Intel aims to redefine the next generation of in-vehicle AI systems, emphasizing the importance of local processing capabilities [40][48]. - The emergence of end-side models is seen as a response to the challenges of cloud-based solutions, positioning them as the future of automotive intelligence [44][46].