Not as strong as DeepSeek V4.1 Flash 1017 on deepswe so it's already dead to me.
Unless they come in at half the price (which I think will be impossible) on inference ... they have wasted their training compute imho.
Not as strong as DeepSeek V4.1 Flash 1017 on deepswe so it's already dead to me.
Unless they come in at half the price (which I think will be impossible) on inference ... they have wasted their training compute imho.
[dead]