About
Quantum Information and Quantum Computation Open Problem Zoo
A curated, citable catalog of open problems in quantum information and quantum computation.
What the zoo is
The QIQCOP Zoo is a place to explore research problems in quantum information and quantum computation. Each problem page brings together a clear statement, the background needed to understand it, key references, and what is known so far.
Browse by field or topic, follow the sources, or share a problem you think belongs here. The collection grows through contributions from the community and review by the maintainers. Problem pages keep permanent links, so you can return to them as the research develops.
A problem page may be out of date or contain errors in its statement, status, progress, or references. Please treat each page as a guide to the literature and check the primary sources and recent work before relying on it or investing effort in a problem. If you find something to correct, tell us.
We would also like to highlight the following projects in the research community:
What will be collected
We collect meaningful, significant unsolved problems in quantum information and quantum computation. Each problem should be formulated precisely in mathematical language, with clear assumptions and an unambiguous criterion for a solution.
A problem is marked Solved when a complete proof or counterexample has been made public in a preprint or a published paper, and its page states whether that result has been peer reviewed. Partial results are recorded as progress and the problem remains Unsolved. The status records what the cited sources establish. The maintainers do not verify proofs. A solved problem stays in the zoo with its resolution and supporting references.
How to contribute
There are three ways to take part. You can propose a problem, contribute through GitHub, or report substantial progress. The maintainers review every contribution, decide what is recorded, and may decline a proposal or an update.
Propose a problem
Use the proposal form to send a problem, its sources, and what is known. No account is needed. State the problem as precisely as you can and name the papers in which it appears.
The maintainers check each proposal against the literature and publish accepted proposals as records with credit to contributors.
Contribute through GitHub
If you are comfortable with Git and TeX, you can add a record yourself in four steps.
- Fork the repository and run
node scripts/new-problem-id.mjs --createto create a problem template with permanent identifiers. - Write the statement, status, source, progress, references, and comment as TeX fragments in the record's fields, following the contribution guide, and choose one or two fields and one to five topics from
database/tags.json. Runnode scripts/migrate-metadata.mjsafter changing the classifications. - Run
node scripts/sync-tex.mjsto write the record's TeX form, thennode site/build.mjs. The build rejects records with missing fields, unknown or miscounted fields and topics, unresolved citations, or unlabeled equations. - Open a pull request.
To correct an existing record or add a reference to it, use the Edit button on the problem's page, which opens the record on GitHub, or open a GitHub issue that describes the change and cites the primary sources.
Report substantial progress
If you have solved a problem listed here or made substantial progress on one, we encourage you to release the work as a preprint, for example on arXiv or Zenodo, or as a published paper, rather than announcing it only in a GitHub issue or comment. A GitHub thread does not appear in arXiv listings or scholarly search engines, can be edited or deleted at any time, and has no persistent identifier such as a DOI or arXiv number. A preprint or paper provides a dated, citable record and opens the work to peer review.
Please report only results that you understand and have checked yourself, and take the time to do so before announcing them. The problems are shared research questions and not a competition. AI tools are welcome, including through the MCP server described below. If they contributed to a result, please say so.
Once the work is available, send us its arXiv identifier or DOI through the Edit button on the problem's page or a GitHub issue. Please link to the work and do not paste a proof into the issue. The maintainers check that the result is relevant and correctly cited and then update the record. Listing a preprint or paper on a problem page does not mean that the maintainers have verified it.
How to cite
Cite the primary sources for any mathematical claim. To cite a problem page for its statement, status, or stable identifier, use the Cite button on that page. To cite the zoo as a whole:
@misc{qiqcop_zoo,
title = {Quantum Information and Quantum Computation Open Problem Zoo (QIQCOP Zoo)},
howpublished = {\url{https://qiqc-op.com/}},
year = {2026},
note = {Accessed 2026-09-24}
}Use the MCP server
Connect your AI assistant through the Model Context Protocol (MCP) to search the zoo, read problem statements and references, and gather the known results and remaining questions for a research session.
Use a client that supports remote MCP servers over Streamable HTTP. Connect with the address below; no download, local setup, or API key is needed to read the catalog.
- Add the server. In your client's MCP or connector settings, add a remote server named
quantum-open-problems. Paste this server URL and choose Streamable HTTP if a transport is requested:https://api.qiqc-op.com/mcp
- Connect your assistant. Save or enable the connection. For clients that accept URL entries in an
mcpServersconfiguration:JSON configuration for clients using
mcpServersAdd this entry to your existing configuration, then reload the client's MCP connection. Some clients use a settings form instead.
{ "mcpServers": { "quantum-open-problems": { "url": "https://api.qiqc-op.com/mcp" } } } - Ask a research question. For example: “Use the quantum-open-problems MCP to find unsolved problems about quantum channel capacity, then summarize one problem's known progress and references.” The assistant can use
search_problems,get_problem,read_problem,list_references, andbuild_context.
The connection reads the current hosted catalog, including newly published problems. If it fails, check the catalog service status and confirm that your client supports remote MCP. The MCP setup and tool guide also covers local clients and, for collaborators with an API key, authenticated research contributions. For direct downloads, the JSON catalog, taxonomy, and agent guide are available.
Licensing and reuse
The software uses Apache-2.0. New original catalog contributions use CC BY 4.0, allowing sharing, adaptation, and commercial use with attribution, license information, and an indication of changes. Contributors retain their copyright.
Cited papers and other third-party material retain their own terms. See the licensing scope and permissions, content license, and retained copyright notices. Cite the primary sources for mathematical results and preserve the supplied contributor credits when reusing licensed text.
Contributions
This project is developed and maintained by Bikun Li, Qicheng Tang, Changhao Li, Chengkai Zhu, Minbo Gao, Zhong-Xia Shang, Bin Cheng, Shihao Ru, and Naixu Guo.
We thank GaugeForge for its financial support of this project.
Mathematics is typeset with MathJax.