loherj

4 ms·

1. 1.▲
   by loherj 1y ago

   Fair point. More benchmarks are definitely good but I’m optimistic that they will show similar results. Anecdotally, I can say that my personal experience with the model is in line with what the benchmarks claim: It’s a bit smarter than R1,
2. 2.▲
   by loherj 1y ago

   Yes. If you look at the diagram that plots the performance vs the amount of output tokens, you can see that R1T2 uses about 1/3 of the output tokens that R1-0528 uses. Keep in mind, the speed improvement doesn’t come from the model run
3. 3.▲
   by loherj 1y ago

   If anybody wants to try it out, it’s up on chutes: https://chutes.ai/app/chute/4fa0c7f5-82f7-59d1-8996-661bb778...
4. 4.▲
   by loherj 1y ago

   Yes and no. Calling us a lab is not quite right, we are a consulting company. But hacking is not just limited to in between placements, everybody has (at least) 2 days per month to do that, regardless of any work for customers. Also, since