Sycophancy (Lechmazur): leaderboard
Tests whether models maintain consistent judgments when presented with opposite first-person perspectives of the same dispute. Measures sycophancy rate - how often a model agrees with both opposing narrators.
Metric: Sycophancy rate % (lower is better). Source: github.com. Status: saturated. 24 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.6 Flash | 0.5 |
| 2 | Gemini 3.5 Flash Lite | 1 |
| 3 | Qwen 3.7 Max (2026-06-08) | 1.5 |
| 4 | Ernie 5.1 | 2 |
| 5 | Qwen 3.7 Flash | 2.5 |
| 6 | MiMo-V2.5-Pro | 2.6 |
| 7 | MiniMax-M3 | 3.6 |
| 8 | Kimi K3 | 4.5 |
| 9 | Gemma 4 31B (Reasoning) | 4.6 |
| 10 | DeepSeek V4 Pro | 5.1 |
| 11 | Qwen 3.7 Plus | 5.1 |
| 12 | Arcee Trinity Large (Thinking) | 18.9 |
| 13 | Mistral Medium 3.5 (High) | 22.4 |
Interactive version: theaggregate.ai/benchmark?slug=sycophancy-lechmazur · How It Works · Data refreshed daily, snapshot 2026-09-05.