Mikael Godal
Commented on New DeepSeek R1 ModelYou can’t ask the model, it doesn’t know. Sometimes the official model even responds that it is OpenAI 4o for example..Commented on New DeepSeek R1 ModelIt’s now live!!Commented on New DeepSeek R1 ModelIt’s very impressive, I requested QwQ and DeepSeek V3 but Nebius seem focused on training rather than adding models.. ?Upvoted New DeepSeek R1 ModelCommented on DeepSeek v3Its on both deep infra and fireworks now, really impressive model!! Especiay the output token size which is basically unlimitedCommented on DeepSeek v3Chat model is here: https://huggingface.co/deepseek-ai/DeepSeek-V3Commented on DeepSeek v3It would make a dramatic difference for coding! But I think it’s too much for Nebius infrastructure to handle, it’s 700GB large!Upvoted DeepSeek v3Commented on LlaMa 3.3It’s been added!Commented on LlaMa 3.3This, in my test on other infra, is that it’s equal to or more often, better than GPT-4o which is seriously impressive.Upvoted LlaMa 3.3Upvoted QwQCreated QwQ