--- title: "KNOWLEDGE ATLAS has raised prices for the third time this year, increasing by another 10%, as domestic large models collectively abandon the price war" type: "News" locale: "en" url: "https://longbridge.com/en/news/281995289.md" description: "Knowledge Atlas raised prices by 10% again on April 8, marking the third price increase this year. The newly released GLM-5.1 model's token price in the Coding scenario is close to Anthropic's Claude Sonnet 4.6. Knowledge Atlas's stock price rose by over 14%, with a market capitalization of HKD 396.1 billion. The GLM-5.1 continues to lead in programming capabilities, particularly surpassing Opus 4.6 in the SWE-bench Pro benchmark test, setting a new global best score. This model can continuously work for up to 8 hours on a single task, with autonomous planning and execution capabilities" datetime: "2026-04-08T07:34:11.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/281995289.md) - [en](https://longbridge.com/en/news/281995289.md) - [zh-HK](https://longbridge.com/zh-HK/news/281995289.md) --- # KNOWLEDGE ATLAS has raised prices for the third time this year, increasing by another 10%, as domestic large models collectively abandon the price war > KNOWLEDGE ATLAS has raised prices again. On April 8, KNOWLEDGE ATLAS officially released and open-sourced its GLM-5.1 model. The model aggregation platform OpenRouter shows that KNOWLEDGE ATLAS GLM has increased prices by another 10%. After the price adjustment, the token price for cache hits in the Coding scenario of GLM-5.1 is close to the level of Claude Sonnet 4.6 under Anthropic. As of the time of writing, KNOWLEDGE ATLAS (2513.HK) has risen over 14% during trading, priced at HKD 888.5, with a market capitalization of HKD 396.1 billion. **KNOWLEDGE ATLAS has raised prices three times recently, with another 10% increase** This is already the third price increase for KNOWLEDGE ATLAS recently. After launching the new generation flagship model GLM-5 on February 12 this year, KNOWLEDGE ATLAS had previously adjusted the pricing for its GLM Coding Plan package, with an overall increase starting at 30%. On March 16, KNOWLEDGE ATLAS released the world's first general large model GLM-5-Turbo, specifically optimized for the "lobster" scenario, and raised the API price for GLM-5-Turbo by 20%. On April 8, the model aggregation platform OpenRouter showed that KNOWLEDGE ATLAS GLM4 has again increased prices by 10%. Currently, the input price for GLM-5 Turbo is USD 1.2 per million, and the output price is USD 4 per million. The newly released GLM-5.1 has an input price of USD 1.26 per million and an output price of USD 3.96 per million. Based on this calculation, GLM-5.1 has increased its input price by 5% compared to the previous generation model, while the output price has slightly decreased by 1%. The newly released GLM-5.1 continues to lead in programming capabilities, achieving the third highest score globally and the first among domestic and open-source models in the comprehensive average score of the three major code evaluation benchmarks: SWE-bench Pro, Terminal-Bench, and NL2Repo. In the SWE-bench Pro benchmark test, which is closest to real software development, the domestic model has surpassed Opus 4.6 for the first time, setting a new global best score. In addition, unlike current models that mainly focus on minute-level interactions, GLM-5.1 can continuously and autonomously work for up to 8 hours on a single task, which has become the biggest highlight of this model release. During task execution, the model can autonomously plan, execute, and test, actively switch strategies when encountering obstacles, self-repair after errors, and ultimately deliver complete engineering-level results. KNOWLEDGE ATLAS recently released its first annual report since going public. The financial report shows that as of March this year, the API calling price of KNOWLEDGE ATLAS has increased by 83% compared to the end of last year, but demand still exceeds supply. KNOWLEDGE ATLAS CEO Zhang Peng stated that there are many vendors in the market providing API services based on computing power, and the fact that customers accept KNOWLEDGE ATLAS's price increase and continue to choose it indicates that the model's capabilities are indeed leading "I believe the essence of long-term pricing is determined by value. Resources that can effectively replace human labor, improve conversion efficiency, and enhance intelligence levels are scarce and valuable. We focus more on the value created per Token and the value it brings to customers. Only when customers recognize this value will they be willing to pay a higher cost to use it. I believe pricing power is still determined by our technological strength and the leading position brought about by long-term trends." **Surging Demand, Domestic Large Models Abandon Price War Collectively** Unlike the price war narrative that permeated domestic large model manufacturers a year ago, this year, domestic model manufacturers have collectively entered a price increase mode. In March, the three major players, Tencent Cloud, Alibaba Cloud, and Baidu Intelligent Cloud, successively announced price adjustments within a week, raising prices for AI computing power services and large model-related products. Among them, some models in Tencent Cloud's Hongyuan series increased by 463.13%, Alibaba Cloud's PingTouGe Zhenwu 810E and other computing card products increased by 5%-34%, and Baidu Intelligent Cloud's AI computing power-related services increased by 5%-30%, with multiple large models that were previously free for public testing simultaneously ending their bonus period and transitioning to commercial billing. This is partly due to the explosive demand brought about by the popularity of various intelligent agents. Previously, informed sources revealed that Alibaba Cloud's AI computing power, storage, and other products had increased by up to 34%, with a major reason being the surge in Token usage. "During the Spring Festival, AI Agent applications exploded, and Alibaba Cloud's MaaS business achieved its highest growth rate in history from January to March this year." MaaS API is a commonly chosen business model among model manufacturers, providing technical services to call cloud-based AI large models through standardized interfaces. In a previous earnings call, KNOWLEDGE ATLAS CEO Zhang Peng mentioned that AI capabilities have now evolved from being usable and playable to truly solving increasingly complex and important problems, making Token API calls and consumption genuinely convert into economic value. The emergence of new application forms such as "lobster," as well as expectations for device-level native intelligence, will also lead to exponential growth in future API and Token consumption. He also stated that the large model industry has developed for several years, and the industry has been looking for a simple and economically powerful business model that can quickly elevate growth trends. The API is a model that converts AI's infrastructure capabilities into economically operational resources. During Tencent's Q4 2025 earnings call, the management responded to the price increases in the domestic cloud computing market, mentioning that the capacity for infrastructure such as CPUs has long been fully booked. Suppliers typically prioritize serving the largest and most stable customers, while smaller enterprises and cloud service providers no longer believe they can reliably obtain supply chain support and have to turn to ultra-large-scale resource service providers. Ultra-large-scale service providers have always operated at low profit margins, and when demand rebounds, the entire industry has little choice but to raise prices. * * * Written by: Nandu N Video Reporter Lin Wenqi ### Related Stocks - [03413.HK](https://longbridge.com/en/quote/03413.HK.md) - [02513.HK](https://longbridge.com/en/quote/02513.HK.md) ## Related News & Research - [The remarkable phenomenon of Chinese AI going global: priced three times higher than domestically, yet still snapped up by foreign programmers.](https://longbridge.com/en/news/283601046.md) - [PharmEasy's parent has a new playbook: Profit first, growth second](https://longbridge.com/en/news/286890424.md) - [China Telecom launches trial AI Token subscription plans starting at $1.4 per month](https://longbridge.com/en/news/286702846.md) - [Can the business model of operators launching AI token packages be viable?](https://longbridge.com/en/news/286700860.md) - [10:03 ETAdAction Launches Prism, an Offer Decisioning API for Personalized Rewarded Monetization](https://longbridge.com/en/news/286429973.md)