Not much info here but this is big news - an actually published open source model that matches O1 from OpenAI - the model has been available behind an api for a few months.
This is r1 not r1-lite-preview. Supposedly r1-lite is a much smaller model, where r1 here is the 600+B MoE, same size as Deepseek v3 they released earlier.
Not much info here but this is big news - an actually published open source model that matches O1 from OpenAI - the model has been available behind an api for a few months.
Here is an article with someone playing with it:
https://www.datacamp.com/blog/deepseek-r1-lite-preview
This is r1 not r1-lite-preview. Supposedly r1-lite is a much smaller model, where r1 here is the 600+B MoE, same size as Deepseek v3 they released earlier.
It’s been reported that DeepSeek R1’s coding capabilities exceed GPT-o1-low and nearly match GPT-o1-meduim, quite astonishing.
If what here says is true: https://x.com/teortaxesTex/status/1880768996225769738, then R1 may as well just be the better model. You can scale up R1 with lower token count to achieve better than o1 high results.