Elimination Game (Lechmazur): leaderboard
20 Questions-style elimination game where models must identify a target from a large set using yes/no questions. TrueSkill ranking from multi-game tournaments across 60+ models.
Metric: TrueSkill μ. Source: github.com. Saturation forecast: Around March 2027. 61 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 (Medium) | 7.52 | #110 (GPT-5.2) |
| 2 | GPT-5 (Medium) | 5.97 | #98 (GPT-5) |
| 3 | GPT-5 Mini (Medium) | 5.73 | #186 (GPT-5 Mini) |
| 4 | Gemini 3 Flash (Preview) | 5.66 | #82 |
| 5 | GPT-4o (Mar 2025) | 5.5 | #330 |
| 6 | DeepSeek R1 0528 | 5.35 | #222 |
| 7 | Grok 4 | 5.15 | #181 |
| 8 | GPT-4.5 (Preview) | 5.12 | #244 |
| 9 | Claude 3.5 Sonnet (20241022) | 5.1 | #291 |
| 10 | Grok 3 | 5.1 | #304 |
| 11 | Gemini 3 Pro (Preview) | 4.89 | #66 |
| 12 | Claude 3.7 Sonnet | 4.78 | #249 |
| 13 | Gemini 2.5 Flash | 4.73 | #245 |
| 14 | Claude Sonnet 4 (Non-reasoning) | 4.64 | #199 (Claude Sonnet 4) |
| 15 | MiniMax-M2 | 4.57 | #317 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=elimination-game-lechmazur · How It Works · Data refreshed daily, snapshot 2026-10-20.