The country's first! Major release by Mooresoft! Launch of domestic intelligent programming services

“Domestic GPU First Stock” Official Announcement of Major News.

Mole-Thread latest announced the launch of the AI Coding Plan intelligent programming service, the first intelligent development solution built on a domestically produced full-function GPU computing base, with a free 30-day trial. Mole-Thread stated that the AI Coding Plan relies on Mole-Thread’s MTT S5000 full-precision computing power, achieving a doubling of computing efficiency through a hardware-software collaborative architecture.

Prior to this, American AI giant OpenAI also released a macOS version of its AI programming agent Codex. Industry analysts believe that 2026 will become the investment year for AI applications, due to continuous improvements in model capabilities, decreasing computing costs, and accelerated monetization of AI applications.

Mole-Thread Announcements

On February 3, Mole-Thread officially launched the AI Coding Plan intelligent programming service.

According to Mole-Thread, as the first intelligent development solution built on a domestically produced full-function GPU computing base, this service is driven by the powerful full-precision computing capability of MTT S5000, integrating a silicon-based flow inference acceleration engine, and incorporating the top-tier code model GLM-4.7. This marks a key breakthrough in domestic chips and large models in the AI Coding field, breaking bottlenecks in domestic computing power in the realm of AI core productivity tools, and demonstrating practical capability to support high-complexity application development.

Mole-Thread states that the newly launched AI Coding Plan relies on Mole-Thread’s MTT S5000 full-precision computing power, achieving a doubling of computing efficiency through a hardware-software collaborative architecture. On this solid foundation, Mole-Thread has partnered with Silicon-Based Flow to develop through efficient operator fusion and extreme framework optimization, significantly reducing response latency while ensuring code generation quality. At the core model level, the system uses GLM-4.7, which ranks first among open-source and domestic models in Code Arena (a professional evaluation system with over a million global users participating in blind testing). This model outperforms peers in functions like code completion and vulnerability detection, even surpassing GPT-5.2.

Currently, Mole-Thread’s AI Coding Plan service has achieved plug-and-play compatibility with several mainstream programming tools such as Claude Code, Cursor, and OpenCode, allowing developers to switch seamlessly across different development environments.

To meet the needs of teams of different sizes, the AI Coding Plan offers tiered packages:

1. Free Trial: 30-day free experience, suitable for lightweight projects;

2. Lite Plan (120 RMB/quarter): three times the usage quota of Claude Pro, supporting medium-frequency development scenarios;

3. Pro Plan (600 RMB/quarter): five times the quota of Lite, designed for complex project architectures;

4. Max Plan (1200 RMB/quarter): prioritized peak traffic guarantee, meeting high-frequency enterprise-level demands.

Mole-Thread states that from now on, developers can visit the product details page to apply for a free trial or upgrade to an enterprise package to achieve maximum development efficiency. As the AGI era accelerates, Mole-Thread will continue to deepen its full-stack layout of “chip + model + toolchain,” helping Chinese developers succeed at the starting line of intelligent transformation.

Impact Level

It is noteworthy that the popularity of AI programming applications is exploding. On February 2, Eastern Time, American AI giant OpenAI also launched a macOS version of its AI programming agent Codex.

It is reported that this new application helps developers manage multiple intelligent agents working in parallel and collaborate on long-running tasks.

Its significant update allows developers using the Codex app not only to perform targeted editing and pairing of individual programming intelligent agents but also to supervise cross-design, build, release, and maintain software lifecycle teams of collaborative intelligent agents.

OpenAI’s official blog mentioned that enterprises and developers are increasingly relying on Codex for end-to-end development. Since the launch of the most powerful intelligent agent programming model GPT5.2-Codex in mid-December, overall usage of Codex has doubled. Over the past month, more than one million developers have used Codex.

OpenAI CEO Sam Altman previously stated that the Codex app is “our most popular product ever.”

All signs indicate that AI applications are accelerating in their breakout. Guojin Securities believes that 2026 will be a critical year for AI applications to move from “technology validation” to “commercial promotion.” The logic of large models serving as both traffic hubs and commercial tools has been realized.

Huaxin Securities states that 2026 will be the “Golden Year” for AI applications, mainly due to three major turning points: first, technological maturity, with models like GPT-5, Gemini 3, and Qwen-Max in 2025 already possessing strong tool invocation, multimodal understanding, and autonomous planning capabilities; second, continuous policy support, with opinions proposing the deep implementation of “Artificial Intelligence+” initiatives and emphasizing the formation of commercial applications; third, market demand resonance, from cost reduction and efficiency improvement on the B-side to widespread adoption on the C-side.

Founder Securities also believes that 2026 will be the investment year for AI applications, due to ongoing improvements in model capabilities, decreasing computing costs, and accelerated monetization processes. Additionally, the tech industry follows a pattern of “hard for three years, soft for three years, and business models for another three years,” with hardware sectors like optical modules and PCBs seeing significant growth from 2022 to 2025.

(Source: Securities Journal)