LuCheng x Huawei Ascend Partner to Launch DeepSeek R1 Series Inference API and Cloud Image Service Based on Domestic Compute | Z News
Unlimited access, free for a limited time!
The open-source release of the DeepSeek-R1 series has sparked global discussion and attention for its outstanding performance and remarkably low development costs.
Luchentech has partnered with Huawei Ascend to jointly launch DeepSeek R1 series inference APIs and cloud image services powered by Ascend compute.
Free Trial of DeepSeek R1 Series Inference API
In this release, Luchentech has built on its self-developed domestic inference engine to successfully optimize the adaptation of DeepSeek R1 series models for Ascend compute. Performance is on par with high-end GPUs, giving developers efficient, flexible, and stable AI inference services. This helps enterprises cut costs and boost efficiency, accelerating the rapid deployment of intelligent business on domestic hardware and software systems.
Domestic Compute Power, Performance Matching High-End GPUs
Breakthrough adaptation for Ascend compute: Luchentech's self-developed inference engine has achieved deep optimization with domestic Huawei Ascend compute, matching the inference performance of high-end GPUs in the industry and contributing to the domestic AI ecosystem.
Luchentech also supports mainstream compute like NV H800: Beyond domestic options, the DeepSeek R1 series inference API supports common compute resources like NV H800, fully unleashing the potential of heterogeneous clusters to meet inference needs across different scenarios.
Inference Demo
Flexible Models, Full Scenario Coverage
The Luchentech Cloud Platform offers a range of flexible options from the full 671B model to efficient distilled smaller models for the DeepSeek R1 series, meeting diverse needs of developers and enterprises. Users can select the right model for their business scenario, achieving the optimal balance between accuracy and efficiency.
Stable API Service, Zero-Barrier Free Out-of-the-Box Use
To meet commercial deployment needs, Luchentech's DeepSeek R1 series inference API provides stable, flexible API service capabilities, ensuring smooth model operation in large-scale production environments and helping enterprises rapidly deploy AI applications. Thanks to its simple, easy-to-use interface, developers can quickly integrate without worrying about underlying details, focusing instead on business innovation.
Currently, the DeepSeek R1 series inference API is available for unlimited free trial for a limited time — all developers are invited to try it out first.
Trial Link:
https://cloud.luchentech.com/maas/modelMarket
Luchentech Cloud Inference Images
For users with high customization needs who want to privately deploy DeepSeek R1 series models, Luchentech also provides open-source-based inference images through Luchentech Cloud. Users can conveniently launch Ascend 910B or NV H800 compute resources with corresponding images to deploy private inference services. This also supports options from the full 671B model to distilled smaller models, with both domestic Ascend and NV H800 compute available for flexible selection.
Inference Demo
For example, after logging into Luchentech Cloud, select an "ASCEND-910B" instance and use the DeepSeek-R1-Distill-Qwen-7B model and image to run this inference query:
How many positive whole-number divisors does 196 have? Put answer in \box{Your answer}
The corresponding output shows the model has engaged in deep thinking.
By selecting the appropriate instance and image, users can further privately deploy the full 671B DeepSeek R1 model.
Online Experience
Free Trial of Luchentech Video Ocean Text-to-Video
Luchentech Video Ocean V2.0 is officially here. This upgrade features major advances in model quality, speed, and functionality — and it's still completely free, so come take advantage. Its innovative technical architecture and user-oriented video generation experience present you with an entirely new digital interactive world.
Online Experience
https://video.luchentech.com/zh-CN
API Access
https://cloud.luchentech.com/maas/modelMarket
Promotional Events
Luchentech Cloud has launched a series of promotional events aimed at delivering real benefits to every user. Whether you're a developer, newcomer to AI, expert, or compute service provider, you can enjoy abundant discounts.
Event Details
https://cloud.luchentech.com/doc/docs/welfare/invitation-bonus
Recommended Reading
