LingOly-TOO: leaderboard
Linguistics reasoning benchmark evaluating models on baseline and obfuscated questions to separate reasoning ability from memorization.
Metric: Obfuscated Score (imported). Source: huggingface.co. Source: benchmarklist.com. Saturation forecast: Around December 2026. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 46.7 | #98 |
| 2 | Claude Opus 4.1 | 45.8 | #140 |
| 3 | Claude 3.7 Sonnet | 42.89 | #249 |
| 4 | Gemini 2.5 Pro | 42.35 | #152 |
| 5 | DeepSeek V3.1 Terminus | 42.2 | #220 |
| 6 | O1 Preview | 32.22 | #193 |
| 7 | O3 Mini (High) | 30.59 | #272 (O3 Mini) |
| 8 | Claude 3.5 Sonnet | 28.1 | #350 |
| 9 | DeepSeek R1 | 26.5 | #251 |
| 10 | GPT-4.5 | 25.45 | #274 |
| 11 | Gemini 1.5 Pro | 20.46 | #445 |
| 12 | GPT-4o | 15.63 | #343 |
| 13 | O3 Mini (Low) | 12.22 | #272 (O3 Mini) |
| 14 | Phi-4 | 11 | #706 |
| 15 | Llama 3.3 70B Instruct | 8.21 | #534 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=lingoly-too · How It Works · Data refreshed daily, snapshot 2026-10-20.