PCB-Bench - Routing Macro QA SBERT: leaderboard
Open-ended question answering about PCB routing, scored by SBERT similarity and macro-averaged over boards.
Metric: SBERT similarity (%). Source: digailab.github.io. Saturation forecast: Not forecast. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 4 Maverick | 51.49 | #464 |
| 2 | GPT-4o | 50.46 | #343 |
| 3 | DeepSeek V3.1 | 49.39 | #264 |
| 4 | Ministral 3B | 49.18 | #1278 |
| 5 | Claude Opus 4.1 | 48.37 | #140 |
| 6 | InternVL3-78B | 46.77 | #352 |
| 7 | MythoMax-L2-13B | 46.76 | #1409 |
| 8 | GPT-5 | 45.88 | #98 |
| 9 | Gemini 2.5 Pro | 44.28 | #152 |
| 10 | Qwen 2.5 7B Instruct | 21.3 | #862 |
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-routing-macro-qa-sbert · How It Works · Data refreshed daily, snapshot 2026-10-20.