The Chinese capital market received significant policy signals on June 17th at the 2026 Lujiazui Forum. The Chairman of the China Securities Regulatory Commission, Wu Qing, announced the expansion of the fifth set of listing criteria for the Shanghai Stock Exchange's Science and Technology Innovation Board (STAR Market) to include the artificial intelligence large model sector. He also explicitly stated support for compliant Hong Kong-listed companies to access the domestic market. On the same day, the Shanghai Stock Exchange released the "Review Guidelines for AI Large Model Enterprises Applying the Fifth Set of STAR Market Listing Standards," providing detailed supporting rules for reviewing AI companies seeking to list. Following the release of this positive news, shares of MINIMAX-WP (00100) closed up 6.63% that day.
Previously, the fifth set of STAR Market criteria was primarily tailored for unprofitable biopharmaceutical companies, with its core purpose being to admit hard-tech enterprises possessing high growth potential despite short-term unprofitability. This new specialized guidance from the Shanghai Stock Exchange customizes rules specifically for the AI large model sector, moving away from simple profitability or revenue size as preconditions. It focuses on a comprehensive evaluation of a company's value across four key dimensions: outstanding technological advancement, possession of phased implementation results, holding of compliant qualifications, and a broad industry growth space. This approach precisely aligns with the industry characteristics of general-purpose large models, which involve high R&D investment and long return cycles.
Concurrently, regulators are encouraging high-quality Hong Kong-listed technology firms to establish a dual primary listing structure for both A-shares and H-shares. This allows companies to leverage both the international financing channels of the Hong Kong market and the domestic industrial capital platform of the A-share market to continuously enhance their computing power reserves and model iteration R&D. It is noteworthy that MINIMAX-WP has already formally engaged China Securities Co., Ltd. to proceed with its A-share IPO tutoring work.
As a leading global general-purpose large model developer, MINIMAX-WP possesses a complete global technological layout. Its latest self-developed M3 large model has achieved scores in multiple core benchmark tests that rival top international products. The company has established differentiated technological barriers in multimodal AI, intelligent agents, and code engineering, earning high recognition within the industry both domestically and internationally, and holding a core position in China's general-purpose large model industry chain.
The M3 model utilizes a self-developed MSA sparse attention architecture, supporting an ultra-long context of up to 1 million tokens. This has reduced the computational cost for long-text inference by over 90%, significantly outperforming peers in processing long documents and large codebases. In third-party authoritative evaluations, the M3 model ranks among the global leaders in multiple metrics. For instance, it scored 59.0% on the code SWE-Bench Pro test, demonstrating the ability to independently fix complete engineering code and replicate experiments from top-tier conference papers. On the intelligent agent multi-tool collaboration test, it scored 74.2%, showing outstanding capability in autonomously completing multi-step complex tasks. The model natively integrates multimodal capabilities for images, text, and short videos, enabling one-stop processing of multi-media content without requiring additional fine-tuning.
Supported by the implementation of products like the M3 series, MINIMAX-WP has established a clear foundation for commercial growth. The company has already accumulated over 300 million global consumer users, with enterprise-level paying customers reaching a scale of millions. Orders continue to expand across multiple sectors including government and enterprises, internet, scientific research, and software development. The pace of global commercial implementation is steadily accelerating, continuously providing cash flow to support the company's R&D investments and computing power expansion. This fully aligns with the STAR Market's fifth set of criteria for reviewing AI companies' technological barriers and commercialization capabilities.
From a capital market perspective, MINIMAX-WP's current valuation does not fully reflect its technological value and growth potential, indicating significant undervaluation. While the company's market capitalization has steadily increased since its Hong Kong listing, the relatively weak liquidity of the Hong Kong market and greater volatility in foreign investor sentiment make it difficult to fully match the recognition domestic capital has for high-quality AI assets. The A-share market has historically accorded higher valuation premiums to hard-tech and leading large model companies, and it aggregates long-term capital from sources such as social security funds, industrial funds, and domestic industrial capital. As the process for returning to the A-share market accelerates, MINIMAX-WP is poised to encounter a significant opportunity for valuation reassessment.
Comments