Terminal-Bench 4.0-AA
Progress Over Time
Interactive timeline showing model performance evolution on Terminal-Bench 4.0-AA
State-of-the-art frontier
Open
Proprietary
Terminal-Bench 4.0-AA Leaderboard
5 models
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  Google | — | — | — | ||
| 2 |  Meta | — | — | — | ||
| 2 | Meta | — | — | — | ||
| 4 | Meta | — | — | — | ||
| 4 | Meta | — | 1.0M | $4.00 / $20.00 |
Notice missing or incorrect data?
What is Terminal-Bench 4.0-AA?
Artificial Analysis reported accuracy on terminal tasks with executable final-container-state verifiers.
Terminal-Bench 4.0-AA is a text benchmark evaluating models on agents, code, and tool calling tasks. LLM Stats tracks 5 models on this benchmark, scored on a 0–1 scale. The current average is 0.6, with the leader at 0.7.
Compare leaders on the best AI for agents, best AI for code and best AI for tool calling leaderboards.
Current leaders
Gemini 4 Carbon from Google currently leads the Terminal-Bench 4.0-AA leaderboard with a score of 0.652 across 5 evaluated AI models.
FAQ
Common questions about the Terminal-Bench 4.0-AA benchmark and leaderboard.