The authors assembled 1,301 expert-curated tasks spanning computer-aided design, engineering, manufacturing and building modeling, plus electronic design automation and 3D visualization.
Frontier AI agents struggle with end-to-end professional engineering workflows
Across 1,301 tasks in six engineering domains, the best evaluated agent scored 44.3, while only 3.6% of multi-software attempts succeeded.
Top University
Hongcheng Gao · Hailong Qu · Yu Lei · Henghui Sun · Haoyang Li · Yipeng Wei · +16 more
Tsinghua University · Zhiman Inc. · Chongqing University · UCAS · Shandong University
Research Digest··3 min read
Gao et al.
Why this paper
From Tsinghua University and 10 others
In one line
Frontier AI agents achieve only 44.3 out of 100 on the EngiWorld benchmark, failing to reliably automate professional engineering tasks.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ✓Reports numbers on named benchmarks
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§