Share your thoughts, 1 month free Claude Pro on usSee more

Social Reasoning on TactfulToM

69.69Accuracy

DeepSeek-R1

Updated 1mo ago

Evaluation Results

Method	Links
DeepSeek-R1 2026.03		69.69
O3 2026.03		68.6
GPT-5 2026.03		65.47
O3 2026.03		65.44
Distill-Llama-70B 2026.03		62.02
LLaMa3.1-70B 2026.03		55.83
Qwen3-32B 2026.03		55.75
LLaMa3.1-70B 2026.03		55.1
GPT-4o 2026.03		52.81
GPT-4o 2026.03		52.67
Qwen3-32B 2026.03		51.93
SocialR1-8B 2026.03		50.79
SocialR1-4B 2026.03		50.01
SocialR1-4B 2026.03		49.41
SocialR1-4B 2026.03		49.36
SocialR1-8B 2026.03		48.72
SocialR1-8B 2026.03		48.6
GPT-5 2026.03		46.7
SocialR1-8B 2026.03		46.67
Qwen3-8B 2026.03		46.55
SocialR1-4B 2026.03		46.09
Qwen3-4B 2026.03		45.84
SocialR1-8B 2026.03		45.58
Qwen3-8B 2026.03		42.92
SocialR1-4B 2026.03		42.65
Qwen3-4B 2026.03		40.31