EMRB - L5 Spectrum Survey and System Design: leaderboard
EMRB level accuracy for surveying the spectrum and designing communication systems such as OFDM configurations, solved by writing code over raw I/Q data (EMRB, arXiv 2608.24086v1, Appendix Table 5). Percent; higher is better; 13 models at temperature 0 with provider default reasoning.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around April 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Grok 4.5 | 39.8 | #54 |
| 2 | Gemini 3.1 Pro (Preview) | 34.8 | #64 |
| 3 | GPT-5.5 | 34.5 | #33 |
| 4 | Gemini 3.5 Flash | 28.3 | #66 |
| 5 | Claude Opus 4.7 | 20.5 | #51 |
| 6 | GLM-5.2 | 20 | #80 |
| 7 | MiniMax-M3 | 19.8 | #151 |
| 8 | DeepSeek V4 Pro | 18.7 | #112 |
| 9 | DeepSeek V4 Flash | 16.5 | #155 |
| 10 | MiMo-V2.5-Pro | 16.3 | #167 |
| 11 | Claude Sonnet 4.6 | 15.8 | #94 |
| 12 | Kimi K2.6 | 12.5 | #113 |
| 13 | Llama 3.3 70B | 2.4 | #578 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=emrb-l5-spectrum-survey-and-system-design · How It Works · Data refreshed daily, snapshot 2026-11-05.