Core Viewpoint - The stock of Zhiyuan surged over 14% in early trading on February 13, reaching a new high since its listing, following the release of its new generation base model GLM-5 on February 12, which is considered the strongest open-source model in the era of intelligent engineering [1][7]. Group 1 - Zhiyuan officially launched the GLM-5 model, marking a transition from the "Vibe Coding" era of writing code snippets to the "intelligent engineering" era of completing system engineering [3][9]. - The GLM-5 model's programming capabilities align with Claude Opus 4.5 and outperform Gemini3 Pro, indicating significant advancements in AI programming technology [3][9]. - The official GLM Coding Plan sold out immediately, marking a historic first for a domestic AI programming model's paid package [3][9].

Core Viewpoint - The release of GLM-5 by Zhipu marks a significant advancement in AI programming capabilities, transitioning from "Vibe Coding" to "Agentic Engineering," allowing AI to independently manage tasks and code development [1][6]. Group 1: Company Performance - Zhipu's stock opened high on February 12, rising over 40% during trading, reaching a historical high with a market capitalization exceeding 170 billion HKD [1][6]. - A-share companies related to Zhipu, such as Capital Online and YK Technology, experienced a 20% limit-up in their stock prices [1][6]. Group 2: Product Development - GLM-5, Zhipu's newly open-sourced flagship model, features a significant upgrade in technical specifications, with parameters increasing from 355 billion in GLM-4.7 to 744 billion, and pre-training data rising from 23 trillion tokens to 28.5 trillion tokens [2][7]. - The new "Slime" framework in GLM-5 supports larger models and complex reinforcement learning tasks, incorporating an asynchronous reinforcement learning algorithm for continuous learning during long-term interactions [2][7]. Group 3: Market Impact - The open-sourcing of GLM-5 democratizes advanced agent and complex programming capabilities, potentially lowering the barriers for enterprises in AI transformation and automation [5][10]. - Zhipu announced a price increase of at least 30% for the GLM Coding Plan, indicating a shift in pricing strategy influenced by international models like OpenAI, moving from a subsidy-based model to one focused on value and premium pricing [5][10].

Core Viewpoint - The article highlights the significant market performance and technological advancements of Zhipu AI, particularly with the launch of its new flagship model GLM-5, which marks a shift in AI programming capabilities from "Vibe Coding" to "Agentic Engineering" [2][6]. Group 1: Market Performance - On February 12, Zhipu AI's stock surged over 40%, reaching a market capitalization of over HKD 170 billion [2]. - A-share companies related to Zhipu AI, such as Shoudu Online and Youkede, also experienced a 20% limit-up [2]. Group 2: Technological Advancements - The GLM-5 model features a substantial upgrade in parameters from 355 billion in GLM-4.7 to 744 billion, and pre-training data increased from 23 trillion tokens to 28.5 trillion tokens [3]. - The new "Slime" framework allows for larger models and complex reinforcement learning tasks, while the integration of the DeepSeek sparse attention mechanism reduces deployment costs without compromising long text performance [3]. Group 3: Competitive Positioning - GLM-5 achieved a state-of-the-art (SOTA) performance in programming and agent tasks, ranking fourth globally and first among open-source models according to the Artificial Analysis Intelligence Index v4.0 [4]. - The model's capabilities are approaching those of Claude Opus 4.5, particularly in complex systems engineering and long-range agent tasks [4]. Group 4: Pricing Strategy - Zhipu AI announced a price increase of at least 30% for its GLM Coding Plan, indicating a shift in pricing logic influenced by overseas models like OpenAI [7]. - The change reflects a transition from a "subsidy for market" approach to a "value for premium" strategy, emphasizing the intrinsic value of foundational models [7].

Core Insights - The company has announced a structural adjustment to the pricing system of the GLM Coding Plan due to strong market demand and increased user engagement [2][4][6] - The price adjustment will take effect on February 12, 2026, with an overall increase starting from 30% [6][5] - Existing subscribers will not see a change in their pricing [6] Pricing Adjustment Details - The first purchase discount will be canceled, while seasonal and annual subscription discounts will remain [6] - The overall price increase for the GLM Coding Plan will start at 30% [6] Product Development and Market Position - The company is increasing its investment in computing power and model optimization to ensure service stability and product capability [2][4] - On February 12, the company officially released its new generation base model, GLM-5, marking a transition in programming capabilities [7] - The company is progressing with its A-share listing plan after completing its Hong Kong IPO on January 8, 2026 [7]

Core Viewpoint - The article highlights the significant stock price increase of Zhipu, which surged over 70% within a week, driven by strong market demand for its GLM Coding Plan and the announcement of a price adjustment for its services [1][3]. Group 1: Stock Performance - On February 12, Zhipu's stock price rose by over 12%, reaching 345 HKD per share, with a market capitalization exceeding 150 billion HKD [1]. - The stock has shown a weekly increase of over 70%, indicating strong investor interest and confidence in the company's growth prospects [1]. Group 2: Price Adjustment Announcement - Zhipu announced a structural adjustment to the pricing of its GLM Coding Plan, effective February 12, 2026, due to strong market demand and increased user engagement [3][6]. - The adjustments include the cancellation of the first purchase discount while retaining quarterly and annual subscription discounts, with an overall price increase starting from 30% [6]. - Existing subscribers will not see a change in their pricing, ensuring customer retention amidst the adjustments [6]. Group 3: Product Development and IPO Progress - Zhipu officially released its new generation base model, GLM-5, marking a transition in programming capabilities from "writing code snippets" to "completing system engineering" [7]. - The company is progressing with its IPO plans, having submitted the third phase of its IPO guidance report to the China Securities Regulatory Commission, indicating ongoing efforts to list on the A-share market after its successful Hong Kong listing [7].

Core Insights - The release of the new base model GLM-5 by Zhiyu Open Source marks a transition from the "Vibe Coding" era of writing code snippets to the "Intelligent Agent Engineering" era focused on completing system engineering [2][3] Group 1: Breakthrough Innovations - GLM-5 features three major breakthrough innovations that position it as a competitive domestic model aligned with the Anthropic Opus series [4] - Performance benchmarks indicate that GLM-5 achieved the highest score among open-source models in authoritative programming tests, closely approaching the performance of the leading Claude Opus 4.5, particularly excelling in complex system engineering and long-term project handling [4] - The model's core technology has been significantly upgraded, with a parameter scale reaching 744 billion, and it innovatively employs an asynchronous reinforcement learning framework called "Slime" along with a sparse attention mechanism, enhancing intelligence while substantially reducing deployment costs [4] Group 2: Development Workflow Transformation - Accompanying the model is the ZCode tool, which allows developers to describe requirements in natural language, enabling the model to automatically decompose tasks and direct multiple agents to concurrently complete coding, debugging, and previewing processes [4] - The ZCode tool also supports remote command of desktop agents via mobile devices, significantly extending productivity [4]

Core Viewpoint - The release of GLM-5 by Zhiyuan marks a significant advancement in the field of open-source models, transitioning from the Vibe Coding era to the Agentic Engineering era, with GLM-5 positioned as a leading model in alignment with the Anthropic Opus series [1] Group 1: Performance - GLM-5 achieves the highest score among open-source models in authoritative programming benchmarks, closely approaching the performance of top-tier models like Claude Opus 4.5 in real programming scenarios, particularly excelling in complex system engineering and long-term projects [1] Group 2: Core Technology - The model features a substantial parameter scale of 744 billion and introduces innovative technologies such as the asynchronous reinforcement learning framework "Slime" and a sparse attention mechanism, significantly enhancing intelligence while reducing deployment costs [1] Group 3: Development Workflow - Accompanying the model is the ZCode tool, which allows developers to describe requirements in natural language, enabling the model to automatically decompose tasks and manage multiple agents to complete coding, debugging, and previewing processes, even allowing remote control via mobile devices [1]