Qwen 2.5 Coder 32B: Is This Best Open Weight Model Better than GPT-4o?

[@jaredcluff5105]

Fantastic video. You answered the question I had about the model perfectly. The model is better than I expected and I might start using it locally as my go to model.

[@engineerprompt]

glad it was helpful.

[@AB-cd5gd]

How good is your pc to handle a 32B model on local

[@HassanAllaham]

Was waiting you to test it.. Thanks for the good content 🌹🌹🌹

[@mc4ndr3]

query the github mass of pull requests... that may provide training data for generally favorable enhancements to software components, including removing bugs, securing software, unifying style, addressing performance problems, and coding most abstractly and flexibly

[@vincentnestler1805]

this is great. Can you give us the text you used to generate the code. I would love to test it on. my machines as well.

[@engineerprompt]

Added to the video description.

[@Username8281]

How does this compare to Claude 3.5 sonnet (new) - not the old one

[@engineerprompt]

I think it's still behind Sonnet but needs further testing

[@aditya_dev30]

Awesome video man. Can you share the prompt used, that is some very good requirement for testing.

[@engineerprompt]

added to the video description.

[@aditya_dev30]

Thanks​@@engineerprompt

[@milesd153]

I'm curious about your hardware specifications.

[@altdoom5205]

I run a threadripper with 16 cores.. and a GTX1080. I am able to run 120B parameter models locally.. anything below 72B runs pretty fast.

[@alexanderkomanov4151]

Thanks a lot!

[@barackobama4552]

Can you test please xAI?

[@engineerprompt]

Yes, planning on doing that.

[@barackobama4552]

@@engineerprompt thank you!