Comparing Claude 4 to previous 3.7 model:
"It's hard for me to quantify, but it feels like you can throw ever more complex challenges at it and it keeps coming up with excellent results."
Comparing Claude 4 to its predecessor (version 3.7), the new model excels at writing "very readable prose" - with many considering it the best in this regard. It also demonstrates strong coding abilities and can create "instant interactive visualisations of data." The author notes it's difficult to quantify the improvement, but suggests the model can handle increasingly complex challenges while maintaining quality results.