PCB-Bench - Placement Macro CQ: leaderboard
Printed-circuit-board component-placement comprehension questions, scored macro-averaged over boards so every design counts equally regardless of how many questions it carries.
Metric: Accuracy (%). Source: digailab.github.io. Saturation forecast: Estimated already saturated. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.1 | 93.3 | #140 |
| 2 | GPT-4o | 92.74 | #343 |
| 3 | DeepSeek V3.1 | 92.74 | #264 |
| 4 | InternVL3-78B | 90.5 | #352 |
| 5 | GPT-5 | 88.27 | #98 |
| 6 | Gemini 2.5 Pro | 86.03 | #152 |
| 7 | Qwen 2.5 7B Instruct | 84.36 | #862 |
| 8 | Llama 4 Maverick | 77.66 | #464 |
| 9 | MythoMax-L2-13B | 65.92 | #1409 |
| 10 | Ministral 3B | 59.21 | #1278 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=pcb-bench-placement-macro-cq · How It Works · Data refreshed daily, snapshot 2026-10-20.