AnswersPublished15 sources
What Lenovo Kaitian’s 1.6-Liter AI Host Offers for Local AI
Lenovo Kaitian’s 1.6 liter AI host combines an eight core Zhaoxin KX 7000 CPU with a Houmo LQ50 NPU rated at 160 TOPS INT8, with a configuration offering up to 320 TOPS. The machine ships with Kylin OS and Tianxi AI Pro, and is marketed for enterprise agent tasks.
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer
Key takeaways
- Lenovo Kaitian’s 1.6 liter AI host combines an eight core Zhaoxin KX 7000 CPU with a Houmo LQ50 NPU rated at 160 TOPS INT8, with a configuration offering up to 320 TOPS.
- The machine ships with Kylin OS and Tianxi AI Pro, and is marketed for enterprise agent tasks.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: What are the key features and reported capabilities of Lenovo Kaitian’s first 1.6-liter AI host for China’s Xinchuang market, unveiled in No. Article summary: Lenovo Kaitian’s first Xinchuang AI host is a compact, 1.6-liter local-AI computer, announced on November 3, 2026, and promoted as a platform for roughly 35-billion-parameter models and an always-on enterprise “digital e. Topic tags: general, general web. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fake numbers, clic
Lenovo Kaitian announced its first Xinchuang-focused AI host on November 3, 2026. The compact 1.6-liter system pairs an eight-core Zhaoxin KX-7000 processor with a Houmo Liqing LQ50 NPU, and is positioned for local inference and enterprise AI-agent tasks. Coverage reports support for models in the 35-billion-parameter class, but its performance figures should be treated as reported results rather than independent benchmarks. 1
2
6
Key specifications at a glance
- CPU: Eight-core Zhaoxin KX-7000.
3
6
- NPU: Houmo Liqing LQ50, with 160 TOPS of INT8 compute in the reported configuration; an NPU configuration of up to 320 TOPS is also described.
1
- NPU memory: 24GB dedicated memory and 153.6GB/s bandwidth.
- Software: Kylin OS and Tianxi AI Pro agent software come preinstalled.
- Size and network: A 1.6-liter chassis weighing about 1.5kg, with two independent 2.5GbE ports, according to product coverage.
2
8
Local model support and reported speed
Reports say the host has been adapted for Qwen3.5 models and Qwen3.6-35B-A3B. One report gives a Qwen3.6-35B-A3B result of 28.99 tokens per second using W4A8 quantization on the 24GB NPU configuration, and says the system supports input contexts of up to 200,000 tokens. These are reported test claims; the available coverage does not provide enough methodology to independently assess or reproduce the results. 2
8
Tianxi AI Pro and the “digital employee” pitch
The host is marketed as a platform for an always-on enterprise “digital employee.” Reports describe Tianxi AI Pro features including document understanding, content generation, information organization, scheduled tasks and skill extensions. 1
3
8
That product positioning is not the same as a verified guarantee about how every task runs or where its data goes. The supplied reports do not establish specific privacy guarantees, cloud-token use, or the full details of a demonstration involving control of a test phone. Those points should be checked against product documentation before being treated as confirmed capabilities.
Cooling, temperatures and noise
The reported cooling design uses separate airflow channels for the CPU and NPU, a hexagonal honeycomb front intake, high-pressure turbine fans and temperature control. 2
3
Coverage reports case temperatures of about 33°C at idle and 41°C under full load in a 25°C room, and about 46°C under full load in a 35°C room. Reported noise levels are 19.4dB at idle and approximately 29.5dB during full NPU inference. These are published figures, not measurements independently verified for this article; the available reports do not establish a consistent test methodology across measurements. 3
6
What the specifications do—and don’t—show
The announced configuration points to a small desktop built to run AI inference locally, with a dedicated NPU, preinstalled agent software and support claims for 35B-class models. The TOPS rating, model compatibility and cited speed are useful starting points, but do not by themselves establish performance for a particular enterprise workload. The available coverage also does not detail Houmo’s quantization, compilation, runtime-management or profiling tools.
For buyers, the key distinction is between the product’s published specifications and its reported demonstrations. Confirm the target models, software support, data-handling behavior and benchmark conditions against vendor documentation and workload-specific tests before relying on them.
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "What Lenovo Kaitian’s 1.6-Liter AI Host Offers for Local AI"?
Lenovo Kaitian’s 1.6 liter AI host combines an eight core Zhaoxin KX 7000 CPU with a Houmo LQ50 NPU rated at 160 TOPS INT8, with a configuration offering up to 320 TOPS.
What are the key points to validate first?
Lenovo Kaitian’s 1.6 liter AI host combines an eight core Zhaoxin KX 7000 CPU with a Houmo LQ50 NPU rated at 160 TOPS INT8, with a configuration offering up to 320 TOPS. The machine ships with Kylin OS and Tianxi AI Pro, and is marketed for enterprise agent tasks.