DeepSeek R1 671B..?

when R1 was added at first it was slow and had context limit issue but it was so smart, after tomorrow the deep seek R1 model become so fast you almost don’t notice it thinking if you turn your face , and the context length has been fixed but since then it became dumber and sometime it do something that you didn’t asked for or can’t understand your request at all , I wonder if its the same first model you guys added or you changed it…

1 Like

it’s the same model! i believe there’s been high load on all inference providers the last couple of days due to dsv3, r1 and o3 being released. popular models!