List your company

Sycophancy detection on SYCON-Bench raw n=100

68Sycophancy RateGemini 3 Flash
-0.6417.183552.82Apr 8, 2026

Evaluation results

MethodLinks
Gemini 3 FlashFamily=Gemini, Organization=Google2026.04680.97
Llama 4 Maverick†Family=Llama, Organization=Meta, Outlier=Discussed in text2026.04580.76
GPT-5.2Family=GPT, Organization=OpenAI2026.04530.72
Gemini 3 ProFamily=Gemini, Organization=Google2026.04530.99
GPT-4.1Family=GPT, Organization=OpenAI2026.04480.83
DeepSeek V3Organization=DeepSeek2026.04470.74
GPT-4oFamily=GPT, Organization=OpenAI2026.04340.65
Llama 4 ScoutFamily=Llama, Organization=Meta2026.04160.9
Sonnet 4.5Family=Claude, Organization=Anthropic2026.0480.88
Opus 4.6Family=Claude, Organization=Anthropic2026.0420.97