BrainBench (EEG) - CodeAct - Sleep Assessment: leaderboard
BrainBench (arXiv 2608.04156v2, Table 1) score on the Sleep Assessment subset built on clinical sleep recordings, with each model running through the free-form CodeAct Python execution loop in an isolated container and its report and artifacts checked by task validators. 0-100; higher is better. Optional reasoning modes disabled.
Metric: EEG analysis score (0-100). Source: arxiv.org. Saturation forecast: Around February 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5 | 73.53 | #16 |
| 2 | GPT-5.6 Sol (Non-reasoning) | 69.24 | #22 (GPT-5.6 Sol) |
| 3 | Qwen 3.7 Max (Non-reasoning) | 67.8 | #85 (Qwen 3.7 Max) |
| 4 | GLM-5 (Non-reasoning) | 64.65 | #149 (GLM-5) |
| 5 | MiniMax-M2.5 | 63.43 | #316 |
| 6 | GPT-5.6 Terra (Non-reasoning) | 62.35 | #40 (GPT-5.6 Terra) |
| 7 | Kimi K2.5 (Non-reasoning) | 62.28 | #152 (Kimi K2.5) |
| 8 | Qwen 3.7 Plus (Non-reasoning) | 62.06 | #98 (Qwen 3.7 Plus) |
| 9 | DeepSeek V4 Flash (Non-reasoning) | 61.01 | #155 (DeepSeek V4 Flash) |
| 10 | Qwen 3.5 122B A10B (Non-reasoning) | 57.18 | #184 (Qwen 3.5 122B A10B) |
| 11 | Qwen 3.5 35B A3B (Non-reasoning) | 50.18 | #268 (Qwen 3.5 35B A3B) |
| 12 | GPT-5.6 Luna (Non-reasoning) | 41.63 | #79 (GPT-5.6 Luna) |
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-codeact-sleep-assessment · How It Works · Data refreshed daily, snapshot 2026-11-05.