Fibble2 Arena (8 guesses): leaderboard
Metric: Daily clean-run win rate (%), 8 guesses. Source: drchangliu.github.io. Saturation forecast: Around January 2027. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 75 |
| 2 | Gemini 2.5 Pro | 50 |
| 3 | GPT-4o | 0 |
| 4 | GPT-5 | 0 |
| 5 | GPT-4o Mini | 0 |
| 6 | Qwen 3 8B | 0 |
| 7 | GPT-5 Mini | 0 |
| 8 | Qwen 3 32B | 0 |
| 9 | DeepSeek V3 | 0 |
| 10 | Qwen 3 14B | 0 |
| 11 | Gemma 3 27B | 0 |
| 12 | Qwen 3 30B A3B | 0 |
Interactive version: theaggregate.ai/benchmark?slug=fibble2-arena-8-guesses · How It Works · Data refreshed daily, snapshot 2026-10-09.