allworld.mediaNews from all over the world
Huawei Cloud releases Harmony Coding LLM tailored for ArkTS
Read more
Pandaily
pandaily.com

Huawei Cloud releases Harmony Coding LLM tailored for ArkTS

Huawei Cloud has updated its coding agent CodeArts for HarmonyOS developers by introducing a large language model for Harmony coding, the CodeArts Harmony agent, and a practice center for Harmony developers. The official product branding is Huawei Cloud / CodeArts.

This model was specifically trained to align with ArkTS and HarmonyOS application patterns, rather than functioning as a general-purpose coding assistant.

According to Huawei Cloud, the Harmony coding model was developed based on over a million source materials related to Harmony and more than 100,000 high-quality Harmony training examples. This data covers UI patterns, components, coding standards, and APIs.

Internal tests reported by InfoQ and LeiPhone demonstrated significant improvements: the error rate in thousand-line code decreased by over 80%, the percentage of successful one-pass compilations increased by over 78%, and token consumption per task decreased by over 20%.

The model is available through CodeArts and the Huawei Cloud MaaS model marketplace for enterprise clients.

Functionality of the CodeArts Harmony Agent

The CodeArts Harmony agent represents a comprehensive process that covers all stages—from requirements analysis and design to code generation, building, and compilation. Huawei Cloud states that the agent integrates DevEco CLI, allowing generated code to be compiled and viewed on HarmonyOS device simulators for phones and tablets directly within the toolset.

Built-in capabilities also include working with Harmony knowledge bases, converting third-party libraries, and migrating mini-programs into atomic services, while documentation and APIs are kept up-to-date.

Alongside the model and agent, Huawei Cloud launched a practice center for Harmony developers. This center combines project examples, code repositories, and browser-based cloud environments, enabling developers to conduct testing, packaging, and compilation without needing a local DevEco installation.

Huawei Cloud notes that the audience for this update exceeds 11 million registered HarmonyOS developers who work on applications from scratch, system migrations, and atomic services.

For Pandaily readers, it is important to note that this news concerns the native Harmony coding product stack, with specific metrics on errors and compilation, rather than financial events. It remains an open question how relevant these internal metrics will remain when working with large production repositories with complex engineering contexts.

Similar stories

Huawei Cloud Announces Commercial Launch Dates for Ascend 950 Lingqu Cluster for China and Global Market
Read more
pandaily.com

Huawei Cloud Announces Commercial Launch Dates for Ascend 950 Lingqu Cluster for China and Global Market

Huawei Cloud CEO Zhou Yu Feng announced at the Huawei Connect 2026 event plans for the commercial launch of a cloud service based on the Ascend 950 Lingqu cluster. The launch in China is scheduled for September 30, and access for the global market will open on November 30.

These dates define the timeline for Ascend 950 compute SKUs and do not merely describe the SuperPoD architecture. Customers will be able to reserve a cluster connected via UnifiedBus consisting of 1024 cards as a managed cloud offering according to the published schedule.

Huawei stated that the Ascend 950 system utilizes its UnifiedBus interconnect, marketed as Lingqu, to assemble a cluster of 1024 accelerators. This cluster is capable of providing up to 1 EFLOPS of computation in FP8 format and 2 EFLOPS in FP4 format, and also features 256 TB of unified memory available globally.

The design is engineered to support end-to-end training of large models without the need for additional cluster partitioning or adaptation, maintaining thousand-card job execution within a single memory address space.

The commercial window for cloud services fits into a broader Ascend development roadmap presented at the same event. Chairman Wang Tao noted that over 1000 Ascend supernodes have already been deployed, and Ascend 950 supernodes have entered a phase of scalable commercial use.

Furthermore, the Ascend 960DT is planned for release in the first quarter of 2027, and the Ascend 960PR in the third quarter, with a stated release pace of approximately one major Ascend generation annually, including long-term stages 970 and 980.

For potential buyers, the key information is the specific calendar with the launch first in China and then globally, as well as the cluster specification: 1024 cards, EFLOPS class throughput (FP8/FP4), and 256 TB of unified memory via UnifiedBus. Huawei Cloud positions Ascend 950 Lingqu as a ready-to-use AI cluster service, rather than a theoretical architecture, offering internal availability at the end of September and international availability two months later for teams needing supernode-scale training without building the infrastructure themselves.

Popular