KORGym - Mathematical and Logical: leaderboard
The "Mathematical and Logical" subtask of KORGym. Knowledge-Orthogonal Reasoning Gymnasium — interactive game-based reasoning benchmark of ~50 games spanning control/interaction, mathematical-logical, puzzle, spatial-geometric and strategic categories, designed to measure reasoning abil...
Metric: Score. Source: razor233.github.io. Saturation forecast: Estimated already saturated. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 0.93 | #123 |
| 2 | Doubao-1.5-Thinking-Pro | 0.84 | #134 |
| 3 | O3 Mini | 0.79 | #272 |
| 4 | DeepSeek R1 | 0.69 | #251 |
| 5 | O1 (2024-12-17) | 0.65 | #149 |
| 6 | Qwen 3 32B (Thinking) | 0.58 | #443 (Qwen 3 32B) |
| 7 | Claude 3.7 Sonnet (Thinking) | 0.52 | #249 (Claude 3.7 Sonnet) |
| 8 | DeepSeek R1 Distill Qwen 32B | 0.35 | #655 |
| 9 | Gemini 2.0 Flash (Thinking) | 0.34 | #342 (Gemini 2.0 Flash) |
| 10 | DeepSeek V3 (0324) | 0.27 | #340 |
| 11 | Gemini 2.0 Flash | 0.17 | #342 |
| 12 | Doubao-1.5-Pro | 0.16 | #287 |
| 13 | GPT-4o | 0.08 | #343 |
| 14 | DeepSeek-R1-Distill-Qwen-7B | 0.06 | #1415 |
| 15 | Qwen 2.5 72B Instruct | 0.04 | #454 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=korgym-mathematical-and-logical · How It Works · Data refreshed daily, snapshot 2026-10-20.