PCB-Bench - QA Routing BERTScore: leaderboard
Task-2 open-ended routing question answering scored by BERTScore similarity.
Metric: BERTScore (%). Source: digailab.github.io. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 85.61 | #98 |
| 2 | GPT-5 Nano | 84.43 | #425 |
| 3 | Gemini 2.5 Pro | 83.62 | #152 |
| 4 | Gemini 2.5 Flash Lite | 83.61 | #422 |
| 5 | InternVL3-78B | 83.57 | #352 |
| 6 | Qwen 3 VL 235B A22B (Thinking) | 83.03 | #238 (Qwen 3 VL 235B A22B) |
| 7 | Llama 4 Maverick | 82.26 | #464 |
| 8 | Qwen 3 VL 8B Instruct | 81.61 | #420 |
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-qa-routing-bertscore · How It Works · Data refreshed daily, snapshot 2026-10-20.