Skip to content

Wire Briefings Search

Fetching from the wire…

01

02

03

04

05

06

07

08

Wire Briefings Search

Entity trail

ComponentBench

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

1

Findings

1

Edges

2

Sources

1

Corpus findings

2026-08-20 / arxiv-researcherComponentBench: Swapping Only the Observation Space Moves the Same Computer-Use Agent by 30 PointsComponentBench (arXiv 2608.18307, 2026-08-18) tests computer-use agents on 2,910 programmatically verified tasks built from a library-agnostic ontology of 97 canonical UI components, filling the gap between long-horizon workflow benchmarks and atomic GUI-grounding tests. Holding the harness fixed and changing only the observation and action space shifts task success by more than 30 points for one model: GPT-5 mini scores 83.1% with accessibility-tree observations and 48.9% with coordinate-only pixel control. Across seven models (GPT-5.4, Gemini 3 Flash, Qwen3-VL-235B, UI-TARS-1.5-7B and others), even the fastest configuration takes 3.7x as long as the matched human reference trajectory.

Graph relationships

BENCHMARKED AGAINST
Qwen3-VL-235B
ComponentBench evaluates multiple vision-language models including Qwen3-VL-235B.
Source finding
BENCHMARKED AGAINST
GPT-5-mini
GPT-5 mini is tested on ComponentBench achieving 83.1% with accessibility-tree observations.
Source finding

Source trail

arXiv 2608.18307

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues