# Kimpton AI

> We build what AI learns from.

Custom evaluations and training environments for frontier AI.

- [Koliseum](https://kimpton.ai/koliseum): Synthetic RL environments built from real-world data.
- [EvalRouter](https://evalrouter.ai): One API to run evaluations against any model or agent.

## Koliseum

We turn real-world data into synthetic RL environments at scale, for frontier labs.

- Find failures: We run your model across many tasks and group its errors into clusters.
- Build environments: Each cluster gets an RL environment with verifiable rewards.
- Check integrity: Before delivery, we test each environment for reward hacks, graders that score wrong, leaked or tampered state, bypassed tool rules, escaped resource limits, and flaky scoring.
- Deliver the record: You receive versioned environments, rewards, and every recorded trajectory.

[Learn more](https://kimpton.ai/koliseum).

### Featured environments

- [Open Trading Floor](https://kimpton.ai/koliseum/trading-floor): View simulated trading results and recorded model decisions.
- [Open Poker Room](https://kimpton.ai/koliseum/poker): Explore model decisions and hand histories from six-max poker.

## EvalRouter

One API to run evaluations against any model or agent. Install it in your terminal, your coding agent, or your code.

Every evaluation version is pinned, so results stay comparable. Runs execute in parallel in isolated sandboxes, with cost shown before you start and results exported as JSON, CSV, or HTML. [Explore EvalRouter](https://evalrouter.ai) or [read the docs](https://evalrouter.ai/developers/docs).

## Work with Kimpton

Describe the capabilities you want to measure or improve. We agree on tasks, scoring, and validation, then build the environment. Your agreed deliverables include versioned tasks, scoring methods, run records, and a validation report.

Selected public environments include [Trading Floor](https://kimpton.ai/koliseum/trading-floor), with separate persistent simulated portfolios, and [Poker Room](https://kimpton.ai/koliseum/poker), with shared play and private cards. Private evaluations are designed around your engagement.

The Kimpton finance workspace remains available to existing customers for research, portfolio context, and trade proposals with evidence. Existing customers can [log in](https://kimpton.ai/chat).

## News

- [Introducing Koliseum](https://kimpton.ai/news/introducing-koliseum)
- [The Kimpton finance workspace](https://kimpton.ai/news/march-2026-dashboards-and-trade-proposals)

## Resources

- [EvalRouter](https://evalrouter.ai)
- [Environment catalog](https://kimpton.ai/koliseum)
- [Documentation](https://kimpton.ai/docs)
- [Full documentation](https://kimpton.ai/llms-full.txt)
- [Agent instructions](https://kimpton.ai/agents.md)
- Evaluation enquiries: info@kimpton.ai