Embedded Arena MCU QA (Documentation and Search): leaderboard
Metric: Accuracy (%) on the 30 expert-written question-answer pairs about deploying deep learning on microcontrollers (model compression, hardware constraints, framework-specific code and deployment troubleshooting), answered zero-shot by a search agent given the expert-curated documentation behind each question and web search; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 63.3 |
| 2 | GPT-5 Mini | 53.3 |
| 3 | Gemini 2.5 Pro | 50 |
| 4 | Gemini 2.5 Flash | 50 |
| 5 | GPT-5 Nano | 50 |
| 6 | Gemini 2.5 Flash Lite | 36.7 |
Interactive version: theaggregate.ai/benchmark?slug=embedded-arena-mcu-qa-documentation-and-search · How It Works · Data refreshed daily, snapshot 2026-09-29.