On August 22, Origin Quantum announced that its in-house quantum computing AI assistant BenXiaoYuan has completed a systematic upgrade and is now officially open-sourced. Two underlying components have been released to developers: the QPanda3 quantum programming skill library and a quantum programming MCP (Model Context Protocol) server.
What It Does
The core move is simple: plugging quantum computing power into the AI coding ecosystem's socket.
- Quantum programming MCP: Based on the standard MCP protocol, it connects Origin Quantum's cloud platform real quantum hardware to the AI ecosystem. Developers who integrate it into AI coding tools like Claude Code or Cursor can complete a full closed loop of "program generation → submission to real quantum hardware execution". In Origin Quantum's own words, it equips AI with "real quantum computer gear".
- QPanda3 skill library: Injects a complete quantum programming knowledge system into AI models, letting AI "learn" to write, debug, and optimize quantum programs like a quantum scientist.
Four Capabilities After the Upgrade
1. Code programming: Writing and debugging quantum programs with QPanda3 and VQNet. 2. Professional Q&A: Integrating foundational concepts, algorithm principles, and product documentation. 3. Application scenario assistant: The first phase covers quantum chemistry computation—molecular energy, potential energy surfaces, and Jordan-Wigner transforms—with quantum machine learning and combinatorial optimization planned next. 4. Deep paper research: Systematically organizing research background, technical routes, experimental methods, and core conclusions.
The entry point is the cloud workbench, whose built-in IDE supports real-time file and circuit diagram viewing—"ask while writing, verify as you code".
Why It Matters
The real significance is not that there is one more assistant, but that it offers a new solution to the usability barrier of quantum computing. Quantum programming has historically had two barriers:
1. Language and abstraction: QPanda3 is not as universally accessible as Python. 2. Hardware access: Real-machine scheduling and circuit compilation are black boxes.
BenXiaoYuan's approach delegates the first layer to large language models (knowledge injection via the skill library) and the second layer to MCP (real hardware becomes a service under a standard protocol). Both layers become usable "just like chatting"—quantum computing power is presented to broad developers in this posture for the first time.
From an industry-position standpoint, this fills in a piece of the domestic quantum software ecosystem. Origin Quantum's earlier achievements were mostly on the offline hardware side (such as the superconducting quantum chip PSE-CZ series), while BenXiaoYuan is a purely software, purely open-source, purely toolchain move. It does not claim compute metrics directly; instead it bets on a smarter path: for quantum computing to go mainstream, first make developers' toolchains frictionless.
Two Questions Worth Following
1. Is quantum chemistry a proxy battleground? The first-phase scenario may be a side-door play for "AI + quantum" general computing—molecular energy calculations can also be done with classical approximations, and the real-machine advantage may only appear at much larger system sizes. This concerns validation of real-hardware value. 2. Will the skill library become a cross-platform asset? After open-sourcing, will community members actually connect the QPanda3 skill library to other quantum cloud platforms (e.g., IBM Quantum)? This concerns whether the ecosystem can gain momentum.