(AWQ VERSION) AGI PUBLIC ACHIEVED

llama.cpp benchmarks(WITHOUT REASON):

ARC Easy: 84.56140351

ARC Challange: 69.23076923

Winogrande: 82.3204

REMEMBER THIS MODEL IS TRAINED TO REASON, WHEN IT STARTS TO REASON ITS REALLY AWESOME

Safetensors

Model size

406B params

Tensor type

I32

BF16

Model tree for Guilherme34/Reason-1-AWQ

Base model

Finetuned

Adapter

Quantized

(2)

this model