List your company

Physical Commonsense Reasoning on PIQA (PIQA Score)

81.1PIQA ScoreLLAMA-4-SCOUT
58.63664.46870.376.132Nov 14, 2025Dec 13, 2025Jan 11, 2026Feb 9, 2026Mar 10, 2026Apr 8, 2026May 7, 2026

Evaluation results

MethodLinks
LLAMA-4-SCOUTPARAMS=109B, Variant=ORIGINAL2025.1181.1
LLAMA-4-SCOUTPARAMS=109B, Variant=FCSD2025.1180.8
LLAMA-4-SCOUTPARAMS=109B, Variant=SFT2025.1180.7
QWEN-3-30B MOEPARAMS=30B, Variant=ORIGINAL2025.1180.5
QWEN-3-30B MOEPARAMS=30B, Variant=FCSD2025.1180.4
DEEPSEEK-V2-LITEPARAMS=16B, Variant=ORIGINAL2025.1180.1
DEEPSEEK-V2-LITEPARAMS=16B, Variant=FCSD2025.1179.9
DEEPSEEK-V2-LITEPARAMS=16B, Variant=SFT2025.1178.2
QWEN-3-30B MOEPARAMS=30B, Variant=SFT2025.1177.8
Echo-180MParams=180M, Tokens=10B, Zero-shot=true2026.0570.9
Transformer-180MParams=180M, Tokens=100B, Zero-shot=true2026.0567.1
Mamba-2-180MParams=180M, Tokens=100B, Zero-shot=true2026.0566.8
Mamba-3-MIMO-180MParams=180M, Tokens=100B, Zero-shot=true2026.0566.7
GDN-180MParams=180M, Tokens=100B, Zero-shot=true2026.0566.3
Mamba-3-SISO-180MParams=180M, Tokens=100B, Zero-shot=true2026.0566.1
Echo-50MParams=50M, Tokens=3B, Zero-shot=true2026.0559.5