LLM2014 Logic 2025-08: leaderboard
LLM2014 monthly logic benchmark for 2025-08, reporting aggregate logical-reasoning performance on the source score scale.
Metric: Median Score. Source: raw.githubusercontent.com. Saturation forecast: Estimated already saturated. 45 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 (High) | 86.57 | #98 (GPT-5) |
| 2 | GPT-5 Mini (High) | 81.69 | #186 (GPT-5 Mini) |
| 3 | Grok 4 | 78.22 | #181 |
| 4 | O3 (High) | 71.31 | #124 (O3) |
| 5 | O4 Mini (High) | 70.22 | #180 (O4 Mini) |
| 6 | GPT-OSS-120B (High) | 57.79 | #341 (GPT-OSS-120B) |
| 7 | Qwen 3 235B A22B 2507 (Thinking) | 56.93 | #260 (Qwen 3 235B A22B 2507) |
| 8 | DeepSeek V3.1 (Thinking) | 56.82 | #264 (DeepSeek V3.1) |
| 9 | Gemini 2.5 Pro | 54.58 | #152 |
| 10 | Seed-1.6 (Thinking) | 52.17 | #265 (Seed-1.6) |
| 11 | DeepSeek R1 0528 | 52.1 | #222 |
| 12 | Claude Sonnet 4 (Thinking) | 51.62 | #199 (Claude Sonnet 4) |
| 13 | Qwen 3 30B A3B 2507 (Thinking) | 49.63 | #376 (Qwen 3 30B A3B 2507) |
| 14 | GLM 4.5 Air (Thinking) | 48.88 | #392 (GLM 4.5 Air) |
| 15 | Qwen 3 235B A22B 2507 | 47.23 | #260 |
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-2025-08 · How It Works · Data refreshed daily, snapshot 2026-10-20.