Using the Models Comparator of OpenAI,
I see no advantage of GPT4 compared to GPT4-Turbo or GPT4o.
Intelligence, context, output are all worse compared to the two others.
The price is even 15 x higher than GPT4o.
The 32k context window versions of the GPT-4 models will be shut off in June 2025. The others remain.
GPT-4 still has unique reference quality not offered by other AI models. While priced at higher per-token rates than latest models coded with efficiency optimizations, you are paying for the raw computation employed, and its particular skills - showing here: