CresOWLve - Russian Exact Match: leaderboard
Metric: Exact-match accuracy (%) after lowercasing, punctuation removal and Unicode normalization, original Russian version, on the 2,061 creative questions of CresOWLve (expert-written What? Where? When? puzzles that need several real-world facts joined by a creative leap), zero-shot; non-thinking models with chain-of-thought prompting, thinking models at the stated effort; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 67.78 |
| 2 | Gemini 3.1 Pro (Preview) (Medium) | 65.79 |
| 3 | Gemini 3.1 Pro (Preview) (Low) | 64.1 |
| 4 | Gemini 3 Flash (Medium) | 51.82 |
| 5 | Gemini 3 Flash (Minimal) | 39.98 |
| 6 | GLM-5 | 28.87 |
| 7 | GPT-5.4 (Medium) | 26.88 |
| 8 | Qwen 3.5 397B A17B | 26.15 |
| 9 | GPT-4.1 | 17.03 |
| 10 | Mistral-Large-3-675B-Instruct-2512 | 15.48 |
| 11 | GPT-5.4 (Non-reasoning) | 15.38 |
| 12 | Qwen 3 235B A22B 2507 Instruct | 13.54 |
| 13 | Qwen 3 235B A22B 2507 (Thinking) | 13.34 |
| 14 | DeepSeek V3.2 (Thinking) | 10.48 |
| 15 | GPT-4.1 Mini | 4.85 |
Interactive version: theaggregate.ai/benchmark?slug=cresowlve-russian-exact-match · How It Works · Data refreshed daily, snapshot 2026-10-07.