An open platform that shows every model call: traces, cost, quality scores and prompt versions. Use it in the cloud or self-host it.
All repositories
Evals & observabilityharbor-framework
terminal-bench-1
A benchmark of hard terminal tasks: a set of assignments and a runner where the agent acts as an admin or developer. Has a leaderboard.
Who it is for
For those testing terminal agents on realistic tasks.
How to start
- Install:
uv tool install terminal-bench(orpip install terminal-bench). - Look at the run options:
tb run --help. - Evaluate your agent with
tb runand submit the result to the leaderboard.
Steps are taken from the README. Check the current version in the repository before running them.
Stars over the last 30 days
Star history is accumulating — the chart appears in a few days.
Author's description
A benchmark for LLMs on complicated tasks in the terminal
More in «Evals & observability»
A long-standing ML platform that now covers LLM apps too: tracing, evaluation, a prompt registry and a model gateway. Plugs into most frameworks.
A command-line tool for testing prompts and agents: describe cases in a config and compare models. Also does red teaming and vulnerability scanning. Plugs into CI.
A platform for debugging and evaluating LLM apps: traces, automated quality checks and dashboards. Supports RAG and agent chains, and can be self-hosted.
OpenAI's framework for evaluating models plus a registry of ready-made benchmarks. A historically important project that shows how eval templates are structured.
A pytest-style framework for testing LLM apps: ready-made metrics for RAG, agents and chatbots, plus trace-based checks. Results can be pushed to a cloud.