From 0-Order Selection to 2-Order Judgment: leaderboard
Multiple-choice reasoning benchmarks face dual challenges: rapid saturation from advancing models and data contamination that undermines static evaluations. Ad-hoc hardening methods (paraphrasing, perturbation) attempt to increase difficulty but sacrifice logical validity for surface complexity, falling short to challenge advanced reasoning models. We present LogiHard, a formal framework that dete
Metric: H-Comb (self-reported). Source: benchmarklist.com. Saturation forecast: Not forecast. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5 | 38.33 | #141 |
| 2 | GPT-5.4 | 31.67 | #79 |
| 3 | Kimi K2.5 | 31.15 | #144 |
| 4 | DeepSeek V4 Pro | 30 | #102 |
| 5 | GLM-4.7 | 30 | #192 |
| 6 | Claude Opus 4.6 | 26.67 | #63 |
| 7 | DeepSeek R1 | 26.67 | #251 |
| 8 | O3 | 25 | #124 |
| 9 | Gemini 3.1 Pro (Preview) | 23.33 | #58 |
| 10 | Qwen 3.5 Plus | 18.33 | #125 |
| 11 | DeepSeek V3.2 | 16.67 | #205 |
| 12 | Qwen 3.6 Plus | 5 | #120 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=from-0-order-selection-to-2-order-judgment · How It Works · Data refreshed daily, snapshot 2026-10-20.