TUA-Bench

What is TUA-Bench?

Evaluates terminal-use agents on 120 real-world tasks across five families (document editing, email, web info, scientific/engineering workflows) using…

Released

2026-06-26

Evaluates

Code & Software, General AI, cs.SE

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.28480

TUA-Bench paper and code

• Paper https://arxiv.org/abs/2606.28480
• Code https://github.com/facebookresearch/TUA-Bench
• Website https://www.tuabench.ai

Which sources cite TUA-Bench?

• TUA-Bench

Related Code & Software, General AI, cs.SE benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A3S-Bench
• AARRI-Bench
• ActBench
• Active-SWE
• ActiveSciBench