Do LLMs Build World Models From Text? A Multil — leaderboard
Metric: L3 (viewpoint) (self-reported). Source: benchmarklist.com. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o | 32.4 |
| 2 | Gemma 3 27B | 28.7 |
| 3 | DeepSeek V4 Flash | 26.4 |
| 4 | Gemini 2.5 Flash | 21.3 |
| 5 | Qwen 2.5 7B Instruct | 18.9 |
| 6 | Qwen 2.5 32B Instruct | 14.7 |
| 7 | qwen3.5-flash | 14.2 |
| 8 | Gemma 2 9B (IT) | 13.7 |
| 9 | Llama 3.1 8B Instruct | 13.3 |
| 10 | Mistral 7B Instruct (v0.3) | 1 |
| 11 | SmolLM2-1.7B-Instruct | 0 |
Interactive version: theaggregate.ai/benchmark?slug=do-llms-build-world-models-from-text-a-multil · How the rankings work · Data refreshed daily, snapshot 2026-07-22.