Agents & Tool Use · 2025
UI-TARS: Pioneering Automated GUI Interaction with Native Agents
An open, end-to-end native GUI agent model that acts purely from screenshots, showing open models can match proprietary computer-use systems on interface-control benchmarks.
Editorial record
Plain-language summary
An open, end-to-end native GUI agent model that acts purely from screenshots, showing open models can match proprietary computer-use systems on interface-control benchmarks.
Knowledge graph
Relationships
Descendants
Conceptual ancestorEvidence: Strongly supported
ReAct: Synergizing Reasoning and Acting in LMs
Builds on the agents lineage in the archive
freshness sweep 2026
Source record
Provenance
- Record ID
- P-600
- Record created
- 2026-07-13
- Last reviewed
- 2026-07-14
- Record version
- 2
- https://arxiv.org/abs/2501.12326
- arXiv:2501.12326
Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.