GPT-4O-Mini + Qwen2 + ContinueDev : This FAST & CHEAP Coding Copilot BEATS Github Copilot & Others!

[@jackflash6377]

That's what I like about this channel.
Always something useful, well explained on how to make it work and typically at the best cost / performance ratio.
Thank you sir!

[@AICodeKing]

Thank you very much!

[@jackflash6377]

@@AICodeKing Welcome. I tried in the past to send Thanks but failed. I finally got it to work so went ahead and made two to make up for the times I tried in the past.
Best channel for real, useful AI knowledge. Thanks again.

[@AICodeKing]

Thanks a lot. I really appreciate the help! I'm glad that my tutorials help people like you. Thanks!

[@orthodox_gentleman]

The Continue dev copilot is absolutely incredible. You can actually execute shell commands via continue extension

[@jayv_tech]

This is what I was looking for. Thanks for sharing.

[@siddharthduggal]

Thanks!

[@ScorgeRudess]

This is insane, thanks!

[@user-jj4ef6dn6z]

Thanks for your work!

[@mrlogo3733]

Thanks for the video.
I will be happy if you will help me with a question:If I add context using chat interface it will use all the data tokens in every prompt?
If I add context using gpts settings and will take it’s (custom gpt’s) api it will work?

[@AICodeKing]

I think it could work.

[@MrMoonsilver]

Hey, what would be the config file if my ollama instance is running on another server in my LAN? Can I just put the API endpoint?

[@AICodeKing]

Yes, you can add an OpenAI model to Continue then in the Config file change the baseURL to Ollama's endpoint.

[@LofiHarmony-gj2by]

I replaced gpt-4o with gpt-4o-mini for my ebook generator. It doesn't work at all. gpt-4o-mini, doesn't even generate outlines.
So, there are big limitations with gpt-4o-mini.

[@jaradaty88]

Can you cover codeuim? Compare it with local models also...it is free and good

[@BadreddineMoon]

Thank you 🎉

[@hsmptg]

Great video (as always)!
Can you give an idea of how many input/output gpt-4o-mini tokens it takes to create a snake game, or a MERN full stack boilerplate for instance?

[@AICodeKing]

For a snake game it could take about $0.1 - $0.2

[@BenAbdallahWalid]

Thanks for the good video.
What do you think what's better QWen or Codestral?
From my point of view Codesral is much better in coding specially if it's built locally

[@AICodeKing]

I prefer Qwen but Codestral can also be good.

[@yoannthomann]

Isn't deepseek better at coding than gpt4o mini

[@AICodeKing]

Yes, much better (if you use their API). Their API is also cheaper than GPT-4O Mini.

[@yoannthomann]

@@AICodeKing it is so difficult to find the best model, do you know a place to compare all of them ?

[@BadreddineMoon]

​@@AICodeKing is deepseek coder v2 api slow just like their free chat?

[@silentspy6980]

​@@AICodeKingis the website capabilities the same as the api ones

[@MrMoonsilver]

It's slow

[@paulyflynn]

We also need local models to debate with each other to come up with the best solution.

[@paulyflynn]

Also, have the local models review your GitHub PRs.

[@AICodeKing]

Sure

[@simion415]

@@paulyflynn Which local model do you prefer to review your Github PRs?

[@brianrowe1152]

What if you have ollama on a different machine locally. So you want to specify the URL instead. ?

[@AICodeKing]

You can do that by specifying baseURL

[@simion415]

@@AICodeKing ​ Can you please share any existing video that you have? How do I use a baseUrl, api_key, api_version and a modelname hosted on a separate machine?

[@birolyildiz]

🎉❤🙏

[@TawnyE]

E

[@paulyflynn]

Thanks!

[@AICodeKing]

Thank you

[@jackflash6377]

Thanks!