AnalogLayoutWorld: An Open Benchmark for Agentic Analog Layout Design
Abstract
Coding agents have demonstrated strong capabilities in software development, yet their ability to tackle long-horizon engineering tasks governed by physical constraints remains less understood. We study this question through analog circuit layout, which demands domain knowledge, geometric reasoning, and iterative interaction with design and verification tools. We introduce AnalogLayoutWorld, an open-source agent benchmark with 40 expert-reviewed tasks for end-to-end layout generation from fixed transistor-level netlists. It combines an integrated design environment with independent physical and post-layout electrical verification, as well as objective layout quality metrics co-designed with analog layout experts. Across 13 model--harness configurations, 91.1% of attempts produce GDS layouts, but only 23.0% satisfy the benchmark's acceptance criteria. We also evaluate Novara, an internally developed agent system, which achieves the highest reported aggregate score. Our analysis links performance differences to test-time scaling under fixed budgets, suggesting that reasoning efficiency and effective use of verification feedback help sustain design quality as task complexity grows.
Figures & tables
| Model / System | Harness | DRC (%) | LVS (%) | SIM (%) | Acc. (%) | Area (%) | Quality | Tool Calls | Toks (M) | Avg. Score |
|---|---|---|---|---|---|---|---|---|---|---|
| Novara Agent System [locally deployed] | 97.5 | 100.0 | 75.0 | 72.5 | 52.2 | 0.816 | — | — | 0.8154 | |
| DeepSeek V4 Pro [max] | Claude Code | 23.3 | 23.3 | 12.5 | 3.3 | 11.3 | 0.049 | 226.9 | 23.8 | 0.1504 |
| DeepSeek V4.1 Flash [max] | Claude Code | 41.7 | 29.2 | 17.5 | 8.3 | 35.0 | 0.086 | 134.0 | 13.2 | 0.1974 |
| DeepSeek V4.1 Flash [max] | DSH | 37.5 | 30.0 | 17.5 | 9.2 | 44.5 | 0.118 | 133.1 | 14.7 | 0.2246 |
| GLM-5.2 FP8 [max] | OpenCode | 10.0 | 10.0 | 5.0 | 0.0 | 21.3 | 0.000 | 242.3 | 18.3 | 0.1075 |
| GLM-5.3 [max] | OpenCode | 80.0 | 69.2 | 42.5 | 35.0 | 35.6 | 0.358 | 264.8 | 34.5 | 0.4779 |
| Submission | GDS | LVS | DRC | Area | Quality | SIM | Total |
|---|---|---|---|---|---|---|---|
| GPT-6 Astra | 0.1000 | 0.1000 | 0.1000 | 0.1651 | 0.0929 | 0.4000 | 0.9580 |
| DeepSeek V4.1 Flash | 0.1000 | 0.1000 | 0.1000 | 0.0956 | 0.0462 | 0.4000 | 0.8418 |
Appendix figures & tables4 assets
Supplementary material from the paper’s appendix.
Appendix
| Task | Task | Task | Task | Task | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 3 | 1.05 | 9 | 4 | 0.60 | 17 | 4 | 0.70 | 25 | 4 | 0.80 | 33 | 3 | 0.80 |
| 2 | 4 | 1.10 | 10 | 3 | 0.90 | 18 | 4 | 0.95 | 26 | 3 | 0.85 | 34 | 3 | 0.95 |
| 3 | 4 | 1.05 | 11 | 3 | 0.90 | 19 | 3 | 0.95 | 27 | 4 | 0.60 | 35 | 3 | 0.95 |
| 4 | 4 | 0.60 | 12 | 3 | 0.95 | 20 | 4 | 0.65 | 28 | 4 | 0.80 | 36 | 3 | 0.95 |
| 5 | 3 | 0.90 | 13 | 4 | 0.65 | 21 | 3 | 0.95 | 29 | 3 | 0.95 | 37 | 4 | 0.90 |
| 6 | 4 | 0.90 | 14 | 4 | 0.90 | 22 | 3 | 1.10 | 30 | 4 | 0.90 | 38 | 3 | 0.80 |
Electronic Design Automation
Electronic Design Automation uses algorithms and software to design, verify, optimize, and manufacture electronic circuits and systems.
See all 244 papers