BrainBench (EEG) - BrainAgent - Overall: leaderboard
BrainBench (arXiv 2608.04156v2, Table 1): LLMs analyse real EEG recordings from 17 datasets (172 tasks) through the structured BrainAgent tool workflow in isolated containers and write reports checked by numerical, categorical, semantic and artifact validators. Overall score across four subsets, 0-100; higher is better. Optional reasoning modes disabled.
Metric: Difficulty-weighted EEG analysis score (0-100). Source: arxiv.org. Saturation forecast: Around May 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5 | 71.95 | #16 |
| 2 | GPT-5.6 Sol (Non-reasoning) | 71.59 | #22 (GPT-5.6 Sol) |
| 3 | Qwen 3.7 Max (Non-reasoning) | 69.77 | #85 (Qwen 3.7 Max) |
| 4 | GPT-5.6 Terra (Non-reasoning) | 67.69 | #40 (GPT-5.6 Terra) |
| 5 | Qwen 3.7 Plus (Non-reasoning) | 67.25 | #98 (Qwen 3.7 Plus) |
| 6 | Kimi K2.5 (Non-reasoning) | 65.08 | #152 (Kimi K2.5) |
| 7 | GPT-5.6 Luna (Non-reasoning) | 63.5 | #79 (GPT-5.6 Luna) |
| 8 | GLM-5 (Non-reasoning) | 63.42 | #149 (GLM-5) |
| 9 | DeepSeek V4 Flash (Non-reasoning) | 63.13 | #155 (DeepSeek V4 Flash) |
| 10 | MiniMax-M2.5 | 60.13 | #316 |
| 11 | Qwen 3.5 122B A10B (Non-reasoning) | 59.25 | #184 (Qwen 3.5 122B A10B) |
| 12 | Qwen 3.5 35B A3B (Non-reasoning) | 52.53 | #268 (Qwen 3.5 35B A3B) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=brainbench-eeg-brainagent-overall · How It Works · Data refreshed daily, snapshot 2026-11-05.