AI Models, Software & Research
Huawei Cloud launched openPangu-2.1-Pro on its model-as-a-service platform, adding capabilities for agents, coding, Ascend operator generation, and inference optimization.
· ComputeLabs Research · from the November 15, 2026 edition
Huawei Cloud says openPangu-2.1-Pro became available on November 15. The model launched on its model-as-a-service (MaaS) platform. The announcement describes it as an open-source Pangu model, although the supplied message does not provide a repository or license.
The capability update is organized around three areas: agent development, AI coding, and inference optimization. Named agent-related scenarios include deep research and HarmonyOS system operation. Coding-related improvements include HarmonyOS application development.
The infrastructure-related capabilities include Ascend operator generation and inference-pipeline optimization. These are stated areas of enhancement rather than quantified performance results. The supplied announcement does not disclose model size, benchmark scores, inference latency, serving cost, or required hardware configurations.
Additional reporting
- Huawei Cloud
- Ascend

