Reportedly: Moore Threads completes Day-0 adaptation of 3-trillion-parameter Kimi-K3.1 Pro, supporting text and image inference
Reported by IT Home · not yet confirmed by the company or a second independent outlet. We update this page when it is.
Moore Threads completed Day-0 adaptation of Moonshot AI's Kimi-K3.1 Pro model on MTT S5000 and MUSA, enabling model loading, service startup, and basic text and image inference, IT Home reported. The adaptation was announced on November 6, 2026, two days after Moonshot AI released the full model weights.
China context
- Original name
- 摩尔线程
- Outside China
- Not stated in the sources yet
- Claims
- Company-reported; not yet independently evaluated
- For builders
- Developers targeting Chinese deployments can evaluate MTT S5000 and MUSA for running Kimi-K3.1 Pro, but must verify performance and compatibility independently.
- For investors
- Moore Threads' rapid support for a leading open-weights model may strengthen its position in China's domestic AI infrastructure market, though revenue impact is unconfirmed.
Translated from Chinese. Quotes and facts link to the original sources.
Moore Threads announced Day-0 support for Moonshot AI's Kimi-K3.1 Pro, a 3-trillion-parameter open-weights model, on its MTT S5000 AI training-inference integrated compute card and MUSA software stack. The adaptation covers model loading, service startup, and basic text and image inference. Moonshot AI released the full model weights on November 4, 2026; Kimi K3.1 Pro (Max) ranked first among open-weights models and 13th overall on Arena's Agent Arena leaderboard, leading in work, coding, and chat categories.
The adaptation uses Moore Threads' MTT S5000 and MUSA software stack, indicating compatibility with a domestic GPU alternative for large-model inference. The evidence does not specify performance metrics, supported precision, or whether the adaptation extends to training workloads.
Chinese AI labs releasing open-weights models gain faster hardware support from domestic chip vendors, reducing dependence on foreign GPUs for inference. Moore Threads' Day-0 adaptation positions its hardware as a viable option for deploying frontier-scale open models in China.
For enterprises in China seeking to deploy large open-weights models without access to restricted foreign hardware, Moore Threads' adaptation offers a potential domestic inference path. The evidence does not confirm availability outside China or commercial terms.
A verifiable next signal is whether Moore Threads publishes benchmark results or deployment guides for Kimi-K3.1 Pro on MTT S5000, or whether other Chinese chip vendors announce similar adaptations.