LLM Stats (Include): leaderboard
197,243 questions from educational and professional exams across 44 languages, measuring region-specific cultural and linguistic knowledge rather than translated English content.
Metric: Score (%). Source: llm-stats.com. Source: qwen.ai. Saturation forecast: Around January 2027. 31 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 87.6 | #37 |
| 2 | Claude Opus 4.6 (Thinking) | 87.4 | #63 (Claude Opus 4.6) |
| 3 | Qwen 3.7 Max | 86.2 | #75 |
| 4 | DeepSeek V4 Pro (Reasoning, Max Effort) | 86.1 | #102 (DeepSeek V4 Pro) |
| 5 | Qwen 3.5 397B A17B | 85.6 | #146 |
| 6 | Qwen 3.6 Plus | 85.1 | #120 |
| 7 | GLM-5.1 (Thinking) | 84.3 | #97 (GLM-5.1) |
| 8 | Kimi K2.6 (Thinking) | 84.2 | #104 (Kimi K2.6) |
| 9 | Qwen 3.7 Plus | 83 | #92 |
| 10 | Qwen 3.5 122B A10B | 82.8 | #173 |
| 11 | Qwen 3.5 27B | 81.6 | #187 |
| 12 | Qwen 3 235B A22B 2507 (Thinking) | 81 | #260 (Qwen 3 235B A22B 2507) |
| 13 | Qwen 3 VL 235B A22B Instruct | 80 | #273 |
| 14 | Qwen 3 VL 235B A22B (Thinking) | 80 | #238 (Qwen 3 VL 235B A22B) |
| 15 | Qwen 3.5 35B A3B | 79.7 | #256 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-include · How It Works · Data refreshed daily, snapshot 2026-10-20.