TextQuests Game Progress
Games · 2025-07-31
TextQuests evaluates LLM agents playing 25 classic interactive-fiction games. This row reports game progress, the percentage progress toward completing all games; higher is better.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Fable 5 | 56.1 |
| GPT-5.6 Sol | 52.1 |
| GPT-5.5 | 48.9 |
| Gemini 3.1 Pro Preview | 45.8 |
| GPT-5.4 | 42.2 |
| Gemini 3 Pro Preview | 41.0 |
| Opus 4.6 | 40.8 |
| GPT-5.6 Terra | 40.5 |
| Opus 4.8 | 40.1 |
| Kimi K3 | 40.0 |
| Opus 4.5 | 38.7 |
| Opus 4.7 | 37.0 |
| GPT-5.2 | 37.0 |
| Gemini 3 Flash Preview | 36.4 |
| Gemini 3.5 Flash | 35.7 |
| GPT-5.1 | 34.2 |
| GLM-5.2 | 33.3 |
| Sonnet 4.6 | 31.5 |
| Sonnet 4.5 | 31.0 |
| o3 | 30.9 |
| GPT-5.6 Luna | 30.8 |
| GPT-5 | 30.0 |
| GPT-5.4 Mini | 29.6 |
| GLM-5.1 | 29.5 |
| Grok 4 | 27.8 |
| Kimi K2.6 | 27.4 |
| Grok 4.5 | 25.4 |
| Muse Spark 1.1 | 24.9 |
| Sonnet 4 | 24.7 |
| Kimi K2.5 | 24.4 |
| Gemini 2.5 Pro | 23.2 |
| DeepSeek-V3.2 | 21.2 |
| DeepSeek-V4-Pro | 20.3 |
| Grok 4 Fast | 20.1 |
| Gemini 3.1 Flash-Lite Preview | 18.8 |
| Grok 4.20 | 18.5 |
| Kimi K2 Thinking | 18.3 |
| GPT-5 Mini | 17.6 |
| Grok 4.1 Fast | 16.4 |
| DeepSeek-R1 | 15.2 |