A-Fantasia
General QA · 2025-04-27
A-Fantasia evaluates mental-imagery-like capability with 300 immediate-answer trials across legal chess moves, 3D cube rotations, and backwards spelling from definitions. Its headline score is the mean error rate across the three 100-item tasks, so lower scores indicate better performance.
Top models (lower is better)
| Model | Score |
|---|---|
| Opus 4.6 | 30.0 |
| GPT-5.5 | 34.7 |
| Opus 4.5 | 35.3 |
| Opus 4.1 | 35.3 |
| GPT-5.6 Sol | 36.0 |
| GPT-4.5 | 37.7 |
| Gemini 3 Flash Preview | 38.3 |
| Gemini 3.1 Pro Preview | 40.0 |
| Sonnet 4.5 | 40.3 |
| Opus 4 | 40.3 |
| Gemini 3 Pro Preview | 41.7 |
| GPT-5.4 | 42.3 |
| GPT-5.6 Terra | 43.3 |
| Sonnet 4 | 44.0 |
| Claude 3.7 Sonnet | 45.3 |
| Claude 3.5 Sonnet (Oct 2024) | 46.0 |
| Qwen3.7-Max | 46.0 |
| Grok 3 | 47.0 |
| GPT-4o | 48.0 |
| Gemini 2.5 Pro | 49.0 |
| Opus 4.8 | 50.0 |
| Opus 3 | 50.3 |
| GPT-5.1 | 51.0 |
| GPT-5.2 | 51.3 |
| GPT-5 Chat (2025-08-07) | 51.7 |
| Gemini 2.0 Flash 001 | 52.3 |
| Sonnet 4.6 | 52.3 |
| GPT-4.1 | 53.0 |
| Gemini 3.1 Flash-Lite | 53.3 |
| Gemini 2.5 Flash | 55.3 |
| DeepSeek-V4-Pro | 56.7 |
| GPT-5.6 Luna | 56.7 |
| Qwen3.6 Plus (2026-04-02) | 58.0 |
| Kimi K2.6 | 60.0 |
| Qwen3.7-Plus | 61.7 |
| Gemini 1.5 Pro | 62.3 |
| Inkling | 64.7 |
| Gemini 2.0 Flash-Lite 001 | 65.0 |
| Qwen3 Max (rolling alias) | 66.0 |
| GLM-5 | 66.7 |