math-olympiad
PackAI & modelsmath-olympiad is a plugin that lets an AI agent solve competition math problems such as IMO, Putnam, and USAMO questions. It checks the agent's proofs step by step using adversarial verification with fresh-context verifiers, catching errors that self-verification misses. When a proof cannot be confirmed, it abstains rather than bluffing.
Unavailable. A Claude Code plugin. It installs in Claude Code, and ahel does not serve packs through its connector yet.
In Claude Code, run:
/plugin install math-olympiad@claude-plugins-officialIt installs from Anthropic's official plugin marketplace.
What your AI can do with it
- Solves olympiad-style competition math (IMO, Putnam, USAMO)
- Verifies proofs with fresh-context adversarial verifiers
- Attacks proofs using specific failure patterns
- Catches errors that self-verification misses
- Abstains when a proof cannot be verified instead of guessing
Getting started
- Have an AI agent setup that supports installing plugins.
- Add the math-olympiad plugin to your agent.
- Give the agent a competition math problem and ask it to produce a proof.
- Review the verified proof or the abstention the agent reports.
Signals
- GitHub stars
- 37k
- Forks
- 4k
- Last commit
- Sep 2026
Questions
- What kinds of problems does it handle?
- Competition math in the style of IMO, Putnam, and USAMO problems.
- How does it check proofs?
- Fresh-context verifiers attack the agent's proofs with specific failure patterns, catching mistakes that self-verification misses.
- What happens when it cannot solve a problem?
- It uses calibrated abstention and says so rather than bluffing an answer.
Advanced
- Item type
- plugin
- Key
math-olympiad-anthropics- Source
- github.com/anthropics/claude-plugins-official