Remix.run Logo
Iolaum 8 hours ago

As long as they are transparent about what quant they serve the model and any other optimization they do that also affects performance of inferred tokens.