Best Coding AI Modal and Why?

Discussions
1

Which one is Best AI Coding Modal and Why? Did you tested? Let’s talk!

o3
grok 4
claude 4 sonnet
claude 4 opus
kimi k2 instruct
gemini 2.5 pro
gpt 4.1
deepseek r1 0528

3 Likes
2

Grok

1 Like
3

From my experience:

o3 > claude 4 opus > gemini 2.5 pro > claude 4 sonnet > gpt 4.1 > deepseek r1 0528

2 Likes
4

I use Gemini Flash because it is Cheap and Good

5

sonnet 4 thinking only, yet to try kimi k2, o3 unusable just like all other openai models, grok hesitates a lot to do things

2 Likes
6

Auto, it uses all the same models, and will fix it’s own mistakes, no need to force a more advanced and more expensive model to make simple edits when a cheaper model can do it faster. Auto is very smart about picking models and is free.

1 Like
7

Claude 4 → Kimi K2 → Gemini 2.5 Pro

2 Likes
8

Are you joking, Right? :sweat_smile:

2 Likes
9

I have auto working on codebases with millions of lines of code, and it is faster and obviously cheaper than just using claude. it even makes less mistakes. but every codebase is different maybe…

10

With a sufficient rule set auto is not bad, but I have like 7000 lines of rules last time i checked.

I have been using o3-Pro for analysis and design, set up the code framework, and then fill it out using sonnet or Opus.

Then debug with Grok4

I make them all work together as a team for me.