CresOWLve: leaderboard

Creative problem-solving on CresOWLve (2026), 2,061 expert-written puzzles from the quiz game What? Where? When?, each needing several real-world facts from different domains joined by a non-obvious creative connection, answered zero-shot in the English version. Accuracy judged by GPT-4o against the reference answer, its explanation and accepted alternatives, in percent; higher is better.

Metric: Accuracy (%) judged by GPT-4o against the reference answer, its explanation and accepted alternatives, English 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 February 2027. 14 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3.1 Pro (Preview)76.03#99
2GPT-5.4Medium53.47#135
3Gemini 3 FlashMedium52.16#175
4GLM-537.17#269
5Qwen 3.5 397B A17B31.25#264
6DeepSeek V3.2Thinking25.47#419
7GPT-4.124.99#505
8Qwen 3 235B A22B 2507 Instruct20.82#616
9Qwen 3 235B A22B 2507 (Thinking)20.82#466
10Mistral Large 317.61#832
11GPT-4.1 Mini13.78#789
12Llama 3.3 70B Instruct10.14#1385
13Command A9.07#1296
14OLMo-2-0325-32B-Instruct5.29#2841

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=cresowlve · How It Works · Data refreshed daily, snapshot 2026-11-21.