LLM2014 Logic 2026-02: leaderboard
Metric: Median Score. Source: raw.githubusercontent.com. 46 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 (Thinking) | 78.02 |
| 2 | GPT-5.2 (High) | 76.82 |
| 3 | Gemini 3.1 Pro (Preview) | 75.02 |
| 4 | Seed 2.0 Pro (High) | 69.55 |
| 5 | Gemini 3 Pro | 68.14 |
| 6 | GPT-5 Mini (High) | 67.84 |
| 7 | Gemini 3 Flash (High) | 66.09 |
| 8 | Kimi K2.5 (Thinking) | 60.75 |
| 9 | Grok 4 | 58.51 |
| 10 | GLM-5 (Thinking) | 55.87 |
| 11 | Qwen 3.5 Plus (Thinking) | 53.8 |
| 12 | Claude Sonnet 4.5 (Thinking) | 52.81 |
| 13 | Claude Opus 4.6 | 51.55 |
| 14 | MiniMax-M2.5 | 49.36 |
| 15 | Grok 4.1 Fast (Reasoning) | 47.79 |
Interactive version: theaggregate.ai/benchmark?slug=llm2014-logic-2026-02 · How It Works · Data refreshed daily, snapshot 2026-09-05.