LLM2014 Logic 2025-11: leaderboard
LLM2014 monthly logic benchmark for 2025-11, reporting aggregate logical-reasoning performance on the source score scale.
Metric: Median Score. Source: raw.githubusercontent.com. Saturation forecast: Around January 2027. 53 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 (High) | 83.75 | #98 (GPT-5) |
| 2 | Gemini 3 Pro | 83.55 | #80 |
| 3 | GPT-5.1 (High) | 79.53 | #135 (GPT-5.1) |
| 4 | GPT-5 Mini (High) | 76.15 | #186 (GPT-5 Mini) |
| 5 | Grok 4 | 69.52 | #181 |
| 6 | Claude Sonnet 4.5 (Thinking) | 66.7 | #142 (Claude Sonnet 4.5) |
| 7 | Kimi K2 (Thinking) | 64.58 | #243 (Kimi K2) |
| 8 | Grok 4.1 Fast (Reasoning) | 62.78 | #217 (Grok 4.1 Fast) |
| 9 | Grok 4 Fast (Reasoning) | 62.72 | #254 (Grok 4 Fast) |
| 10 | Qwen 3 Max (Thinking) | 53.6 | #208 (Qwen 3 Max) |
| 11 | Qwen 3 235B A22B 2507 (Thinking) | 51.81 | #260 (Qwen 3 235B A22B 2507) |
| 12 | DeepSeek V3.2 (Thinking) | 50.77 | #205 (DeepSeek V3.2) |
| 13 | Gemini 2.5 Pro | 47.67 | #152 |
| 14 | GPT-OSS-120B (High) | 47.57 | #341 (GPT-OSS-120B) |
| 15 | Qwen 3 Max | 46.94 | #208 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm2014-logic-2025-11 · How It Works · Data refreshed daily, snapshot 2026-10-20.