DeepSeek R1 is far better than OpenAI o1 at finding bugs in pull requests (greptile.com) 43 points by dakshgupta 1y ago ↗ HN
[–] buyucu 1y ago ↗ I hope this moves the masses from closed models to open models. [–] dakshgupta 1y ago ↗ The only reason we haven't moved yet is that there are no serverless inference providers that are has performant as Anthropic/OpenAI (would love to be corrected if I am wrong). [–] buyucu 1y ago ↗ give deepseek a few weeks to catch up. they probably got waaay more customers than they originally anticipated. [–] buyucu 1y ago ↗ looks like azure supports deepseek now: https://azure.microsoft.com/en-us/blog/deepseek-r1-is-now-av... [–] maeil 1y ago ↗ Azure seems to be the least performant of all. [–] buyucu 1y ago ↗ I'm not suprised [–] jedberg 1y ago ↗ Check out LambdaLabs Inference:https://lambdalabs.com/inferenceIt's pretty quick and has the open models. They have Qwen right now, I wouldn't be surprised if they added Deepseek soon.
[–] dakshgupta 1y ago ↗ The only reason we haven't moved yet is that there are no serverless inference providers that are has performant as Anthropic/OpenAI (would love to be corrected if I am wrong). [–] buyucu 1y ago ↗ give deepseek a few weeks to catch up. they probably got waaay more customers than they originally anticipated. [–] buyucu 1y ago ↗ looks like azure supports deepseek now: https://azure.microsoft.com/en-us/blog/deepseek-r1-is-now-av... [–] maeil 1y ago ↗ Azure seems to be the least performant of all. [–] buyucu 1y ago ↗ I'm not suprised [–] jedberg 1y ago ↗ Check out LambdaLabs Inference:https://lambdalabs.com/inferenceIt's pretty quick and has the open models. They have Qwen right now, I wouldn't be surprised if they added Deepseek soon.
[–] buyucu 1y ago ↗ give deepseek a few weeks to catch up. they probably got waaay more customers than they originally anticipated.
[–] buyucu 1y ago ↗ looks like azure supports deepseek now: https://azure.microsoft.com/en-us/blog/deepseek-r1-is-now-av... [–] maeil 1y ago ↗ Azure seems to be the least performant of all. [–] buyucu 1y ago ↗ I'm not suprised
[–] maeil 1y ago ↗ Azure seems to be the least performant of all. [–] buyucu 1y ago ↗ I'm not suprised
[–] jedberg 1y ago ↗ Check out LambdaLabs Inference:https://lambdalabs.com/inferenceIt's pretty quick and has the open models. They have Qwen right now, I wouldn't be surprised if they added Deepseek soon.
9 comments
[ 3.1 ms ] story [ 35.8 ms ] threadhttps://lambdalabs.com/inference
It's pretty quick and has the open models. They have Qwen right now, I wouldn't be surprised if they added Deepseek soon.