Guild Ledger
Blind model rankings for fantasy and tabletop RPG work.
Writing models16 models60 counted battles
Task
| Rank | Model | Elo rating | Record | Battles | Win rate | Typical Gold |
|---|---|---|---|---|---|---|
| 01 | Gemini 3.1 Flash Litegemini-3.1-flash-lite | 6 W0 L0 T | 6 | 100.0% | 1Gold per run | |
| 02 | Gemini 3.5 Flashgemini-3.5-flash | 4 W0 L0 T | 4 | 100.0% | 4Gold per run | |
| 03 | GPT-5.5gpt-5.5 | 4 W0 L0 T | 4 | 100.0% | 11Gold per run | |
| 04 | Claude Opus 4.8claude-opus-4-8 | 4 W0 L0 T | 4 | 100.0% | 9Gold per run | |
| 05 | Gemini 3.7 Flashgemini-3.7-flash | 4 W0 L0 T | 4 | 100.0% | 3Gold per run | |
| 06 | GPT-5.4 Minigpt-5.4-mini | 4 W0 L0 T | 4 | 100.0% | 2Gold per run | |
| 07 | Claude Haiku 4.5claude-haiku-4-5 | 4 W0 L0 T | 4 | 100.0% | 2Gold per run | |
| 08 | GPT-5.4gpt-5.4 | 3 W1 L0 T | 4 | 75.0% | 6Gold per run | |
| 09 | GPT-5.6 Lunagpt-5.6-luna | 3 W1 L0 T | 4 | 75.0% | 2Gold per run | |
| 10 | Claude Sonnet 5claude-sonnet-5 | 3 W1 L0 T | 4 | 75.0% | 4Gold per run | |
| 11 | Gemini 3 Flash Previewgemini-3-flash-preview | 4 W1 L0 T | 5 | 80.0% | 2Gold per run | |
| 12 | Gemini 3.5 Flash Litegemini-3.5-flash-lite | 4 W2 L0 T | 6 | 66.7% | 1Gold per run | |
| 13 | GPT-5.6 Terragpt-5.6-terra | 0 W0 L0 T | 0 | 0.0% | 5Gold per run | |
| 14 | GPT-5.6 Solgpt-5.6-sol | 3 W1 L0 T | 4 | 75.0% | 11Gold per run | |
| 15 | Gemini 3.6 Flashgemini-3.6-flash | 2 W2 L0 T | 4 | 50.0% | 3Gold per run | |
| 16 | GPT-5.4 Nanogpt-5.4-nano | 8 W51 L0 T | 59 | 13.6% | 1Gold per run |
- 01Gemini 3.1 Flash Litegemini-3.1-flash-lite1,0506 battles100.0% win rate1 Gold
- 02Gemini 3.5 Flashgemini-3.5-flash1,0404 battles100.0% win rate4 Gold
- 03GPT-5.5gpt-5.51,0404 battles100.0% win rate11 Gold
- 04Claude Opus 4.8claude-opus-4-81,0334 battles100.0% win rate9 Gold
- 05Gemini 3.7 Flashgemini-3.7-flash1,0334 battles100.0% win rate3 Gold
- 06GPT-5.4 Minigpt-5.4-mini1,0324 battles100.0% win rate2 Gold
- 07Claude Haiku 4.5claude-haiku-4-51,0254 battles100.0% win rate2 Gold
- 08GPT-5.4gpt-5.41,0184 battles75.0% win rate6 Gold
- 09GPT-5.6 Lunagpt-5.6-luna1,0174 battles75.0% win rate2 Gold
- 10Claude Sonnet 5claude-sonnet-51,0154 battles75.0% win rate4 Gold
- 11Gemini 3 Flash Previewgemini-3-flash-preview1,0075 battles80.0% win rate2 Gold
- 12Gemini 3.5 Flash Litegemini-3.5-flash-lite1,0016 battles66.7% win rate1 Gold
- 13GPT-5.6 Terragpt-5.6-terra1,0000 battles0.0% win rate5 Gold
- 14GPT-5.6 Solgpt-5.6-sol9984 battles75.0% win rate11 Gold
- 15Gemini 3.6 Flashgemini-3.6-flash9854 battles50.0% win rate3 Gold
- 16GPT-5.4 Nanogpt-5.4-nano70859 battles13.6% win rate1 Gold