LLM2014 Logic 2026-08: leaderboard
Metric: Median Score. Source: raw.githubusercontent.com. 46 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 (xHigh) | 73.89 |
| 2 | GPT-5.6 Sol (xHigh) | 69.42 |
| 3 | Kimi K3 (Max) | 67.66 |
| 4 | Claude Opus 5 (xHigh) | 64.74 |
| 5 | GLM-5.3 (Max) | 63.91 |
| 6 | GLM-5.3 Flash (Max) | 60.52 |
| 7 | Qwen 3.8 Max (xHigh) | 60.05 |
| 8 | DeepSeek V4 Pro (0813) (Max) | 59.63 |
| 9 | Gemini 3.7 Flash (High) | 57.56 |
| 10 | Grok 4.6 (High) | 56.41 |
| 11 | Gemini 3.1 Pro (Preview) (High) | 55.84 |
| 12 | DeepSeek V4 Flash (0731) (Max) | 55.23 |
| 13 | GLM-5.2 (Max) | 54.7 |
| 14 | GPT-5.6 Luna (xHigh) | 51.62 |
| 15 | Qwen 3.8 27B (xHigh) | 47.65 |
Interactive version: theaggregate.ai/benchmark?slug=llm2014-logic-2026-08 · How It Works · Data refreshed daily, snapshot 2026-09-05.