cluade 4 is leagues better than gpt 4.1 after using 4.1 during the outage its limitations compared to claude 4 were glaring.
I agree in many cases GPT 4.1 seems too creative where Claude 4 Sonnet is more precise.
I tried GPT 4.1 for a day and all it did was coming back to me with my initial plan and told me how to implement it 
1 Like
claude 4 is able to work with huge files and complex project structure with close supervision. gpt 4.1 was barely able to recognize what i was doing and gave up almost instantly. gpt 4.1 is super ‘lazy’ by comparison.