Frequently asked questions
Short answers. Words in bold are explained in the Glossary. Your question is not here? Ask in Discussions.
About the project
What is BuddhistAI? An open project that tests AI models and skills on real Buddhist tasks and shares all results for free. For each task you see the Top performance and the Best value. People who build AI solutions can send theirs to be tested.
Who runs it? It is a shared project. Many people and organizations contribute; none of them owns the brand. How decisions are made is on the Governance page. Contributors and supporters are listed on the homepage.
Is it only about texts? No. Texts are one area. We cover Buddhist tasks in general: translation and study, teaching, art and images, practice support, running centers and projects, and personal AI setups.
Which traditions and languages? All of them. Today most benchmarks are Tibetan and Thai, because that is where people started. See Languages for what exists in each language.
Is it free? Yes. Everything is public and MIT licensed (imported data keeps its own license). Evaluations are paid for by supporters, so testing your solution is free too.
Using the results
Which AI should I use for my task? Open the leaderboard, find your task and language, and choose the Top performance (best quality) or the Best value (good quality, much cheaper). For translation into a Buddhist language, see Languages.
How do I install a skill?
Skills are in buddhist-skills. In Claude Code: claude plugin marketplace add BuddhistAI/buddhist-skills, then claude plugin install buddhist-skills-<language>@buddhistai (for example buddhist-skills-tibetan). You can also copy the SKILL.md file into any assistant that reads skills, or paste it as instructions.
Can I trust the scores? Each score says which skill version, model and data version it comes from, and has a provenance record with the exact prompt, settings, logs and cost. Anyone can re-score the stored answers (Tutorial 1). Also check the benchmark’s maturity level: a Draft benchmark is less certain than a Reviewed one.
Why is there no benchmark for my task? Nobody has built one yet. You can propose one, or request an AI solution for your task.
Why do costs say “estimated”? Some early runs did not log token counts, so they were estimated from the text length. New runs use the real counts from the API.
Contributing
I am not a programmer. Can I help? Yes. You can share how your project works (a workflow), check answers in your language, review a benchmark as an expert, suggest a task, add a project to the Directory, or fix a page. Most of this is a form or the “Edit this page” link. See Contribute.
How do I submit my model or skill? Paste a link on the homepage. For a skill or any file you give to AI, you can also use the Share your files form. A maintainer runs it on the benchmark. To learn the steps yourself, follow the tutorials.
Can my organization donate API credits or compute? Yes, please. Use the Donate compute or API credits form. Donations pay for evaluations of everyone’s submissions. They never change a score or a ranking.
Our team wants to contribute regularly. Do we get access? Yes. Your organization can get its own GitHub team with write access to your own task repos. The steps are in Governance.
Can I submit a closed or commercial tool? Yes. Include its outputs on the benchmark and say in your submission how others can get the tool. It appears on the leaderboard like any other entry.
Who owns what I contribute? You keep your copyright. Contributions to our repos are published under the MIT license (data under the license you choose and state). You are credited by name.
My skill won. What happens? It becomes the best skill for its task and language in buddhist-skills, we congratulate you, and we announce it in News, crediting you and your organization.

BuddhistAI Knowledge base