WearableQA - Cross-Signal: leaderboard
WearableQA (arXiv 2609.05405v1): 4,084 ten-option multiple-choice questions over wearable time series, blood biomarkers and demographics of 200 real users with up to 500 days of data. Accuracy on questions that integrate two or more signals, chain-of-thought (Table 1). Higher is better.
Metric: Accuracy (%; 10-option multiple choice, chain-of-thought). Source: arxiv.org. Saturation forecast: Around May 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 73.2 | #64 |
| 2 | Claude Opus 4.6 | 56.8 | #67 |
| 3 | Claude Sonnet 4 | 48.5 | #207 |
| 4 | GPT-5.4 | 45.7 | #83 |
| 5 | Gemini 2.5 Pro | 45.3 | #159 |
| 6 | Gemma 4 26B A4B | 43.5 | #360 |
| 7 | GPT-4o | 36.4 | #357 |
| 8 | Mistral Small 3.1 | 30.3 | #623 |
| 9 | Llama 3.3 70B | 29.5 | #578 |
| 10 | Gemma 3 12B | 27.4 | #674 |
| 11 | Llama 3.1 8B | 25.7 | #1130 |
| 12 | Gemma 3 27B | 25.6 | #606 |
| 13 | Llama 3.2 3B | 22.1 | #1376 |
| 14 | Gemma 3 4B | 20.2 | #1088 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=wearableqa-cross-signal · How It Works · Data refreshed daily, snapshot 2026-11-05.