China's Quantum Platforms
◐ Design-level API
This page documents qalgora-Q API design, architecture, or adaptation workflows. Code examples illustrate intended usage and are not guaranteed to run in the current reference implementation.China fields world-leading quantum hardware across photonic and superconducting modalities. qalgora-Q documents intended adapters / export workflows for the major Chinese platforms; the open reference build runs locally and does not submit to these machines directly.
The landscape
| Platform | Builder | Modality | Highlight |
|---|---|---|---|
| Jiuzhang (九章) | USTC · Pan Jianwei | Photonic (GBS) | Jiuzhang 4.0 (2026): 3050 detected photons |
| Zuchongzhi 3.0 (祖冲之) | USTC | Superconducting | 105 qubits; ~15 orders faster than the best classical supercomputer |
| Origin Wukong (本源悟空) | Origin Quantum | Superconducting | Wukong-180 (2026): 180 qubits |
| TuringQ (图灵量子) | TuringQ · Jin Xianmin | Photonic (integrated) | Full-stack; Gen3 targets ~500 modes |
| QuantumCTek (国盾量子) | QuantumCTek · Hefei | Superconducting (cloud) + QKD | Tianyan (天衍) cloud — 176-qubit Zuchongzhi-line chip + 504-qubit "Xiaohong" (骁鸿) chip |
Targeting Chinese hardware
The three paradigms are driven differently. Gate-model superconducting kernels validate locally on the CPU simulator and submit to a superconducting cloud through a planned adapter:
import qalgora
# Gate-model superconducting: Origin Wukong, and the Tianyan / Zuchongzhi cloud.
qalgora.set_target("qpp-cpu") # runnable local CPU statevector validation
counts = qalgora.sample(bell, shots_count=1000)
# Planned adapter — submit the same gate-model kernel to a superconducting cloud:
# qalgora.set_target("origin", machine="wukong") # Origin Wukong-180 (planned)Reference build runs locally — not real hardware
This example targets a remote cloud/QPU, but the open reference build does not submit to a real machine — it warns and falls back to the local CPU statevector simulator, returning simulated results. Submitting to an actual device (e.g. 天衍/TianYan, IBM, IonQ, Origin) requires the vendor’s own SDK and credentials.Photonic systems are not gate-model and do not use sample(bell).
Jiuzhang (Gaussian Boson Sampling) and TuringQ (integrated-photonic interferometry) are addressed
through a separate conceptual photonic API:
# Conceptual photonic API only; not in the open reference implementation.
# Photonic GBS sampling (Jiuzhang) — a specialized sampler, not a gate-model QPU:
# samples = qalgora.photonic.run(gbs_program, target="jiuzhang")
# Integrated-photonic programmable interferometry (TuringQ):
# result = qalgora.photonic.run(interferometer_program, target="turingq")Specification API — not in the open reference build yet
This example shows a qalgora-Q specification API (or a third-party library) that the open reference build does not bundle today. It documents the intended interface; to run code now, use the reference build’s supported core API.Three distinct paradigms
These platforms span three different models: (1) superconducting gate model
(Origin Wukong; the Tianyan / Zuchongzhi cloud) runs gate-based kernels; (2) photonic
Gaussian Boson Sampling (Jiuzhang) is a specialized sampling machine, not a gate-model
QPU; (3) integrated-photonic programmable interferometry (TuringQ) is also
photonic, not gate-model. The two photonic families use a conceptual qalgora.photonic.*
API, never sample(bell).
中国量子平台总览
◐ 设计接口
本页描述的是 qalgora-Q 的接口设计、架构设计或适配工作流。相关代码用于说明预期用法，当前参考实现不保证可以直接运行。中国在光量子与超导两条技术路线上都拥有世界领先的量子硬件。qalgora-Q 记录了面向主要中国平台的预期适配与导出工作流；开放参考实现仅在本地运行，不会直接提交到这些机器。
整体格局
| 平台 | 研制方 | 技术路线 | 亮点 |
|---|---|---|---|
| 九章 | 中国科大 · 潘建伟 | 光量子（高斯玻色采样） | 九章四号（2026）：3050 个探测光子 |
| 祖冲之三号 | 中国科大 | 超导 | 105 比特，比最快经典超算快约 15 个数量级 |
| 本源悟空 | 本源量子 | 超导 | 悟空-180（2026）：180 比特 |
| 图灵量子 | 图灵量子 · 金贤敏 | 光量子（集成光子） | 全栈光量子；Gen3 目标约 500 模 |
| 国盾量子 | 国盾量子 · 合肥 | 超导（云平台）+ QKD | 天衍云 —— 176 比特祖冲之系列芯片 + 504 比特"骁鸿"芯片 |
对接中国硬件
三种范式的对接方式各不相同。门模型超导内核先在本地 CPU 模拟器上验证，再通过规划中的适配器提交到超导云：
import qalgora
# 门模型超导：本源悟空，以及天衍 / 祖冲之云。
qalgora.set_target("qpp-cpu") # 可运行：本地 CPU 态矢量验证
counts = qalgora.sample(bell, shots_count=1000)
# 规划中的适配器 —— 把同一门模型内核提交到超导云：
# qalgora.set_target("origin", machine="wukong") # 本源悟空-180（规划中）参考实现仅本地运行 · 非真机
此示例虽指向远程云端/QPU，但开放参考实现不会真正提交到真机 —— 它会给出告警并回退到本地 CPU 态矢量模拟器，返回模拟结果，不会真正提交到天衍等真机。真正提交真机需使用对应厂商的 SDK 与凭证。光量子系统并非门模型，也不使用 sample(bell)。九章（高斯玻色采样）与图灵量子（集成光子干涉）通过另一套概念性光子接口对接：
# 仅为概念性光子接口；开源参考实现尚未包含。
# 光量子 GBS 采样（九章）—— 专用采样机，并非门模型 QPU：
# samples = qalgora.photonic.run(gbs_program, target="jiuzhang")
# 集成光子可编程干涉（图灵量子）：
# result = qalgora.photonic.run(interferometer_program, target="turingq")规范接口 · 参考实现暂未包含
此示例展示的是 qalgora-Q 规范中的接口（或第三方库），开放参考实现目前尚未内置，仅用于说明预期用法；如需立即运行，请使用参考实现已支持的核心 API。三种不同范式
这些平台涵盖三种不同模型：（1）超导门模型（本源悟空；天衍 / 祖冲之云）运行基于门的内核；（2）光量子高斯玻色采样（九章）是专用采样机，并非门模型 QPU；（3）集成光子可编程干涉（图灵量子）同样是光量子而非门模型。两类光量子系统使用概念性的 qalgora.photonic.* 接口，绝不使用 sample(bell)。