I am currently using TaskingAI + OpenAI’s GPT-4 to build an application embedded in a social app for use as a bot. This process is indeed very convenient, and for some reasons, I highly value the diversity of choices in underlying models. Are there any other models that can compete with OpenAI GPT-4 in terms of cost, performance, or behavior? I have also tried Google’s Gemini, but I think there is still a significant gap. Thank you for your suggestions!
It’s undeniable that GPT-4 remains the most practical LLM available today, but you might want to try Groq’s model, which is incredibly fast.
You can try Moonshot, which boasts an advantage in handling extremely long contexts.
Anthropic’s Claude 3 is quite a bit better than GPT-4 Turbo.