Choosing the Right AI Model: Anthropic vs. ChatGPT and Others

·

I use Clay to help me generate outbound blurbs. I'm currently using Anthropic (Claude 3 Opus specifically), but I'm wondering if I'm using the right model. Do folks have any strong opinions about using Anthropic vs. ChatGPT, Cohere, Gemini or Mistral? Thanks for the guidance and feedback, I appreciate it

6 comments

Enzo G.
·
·
If it's for writing emails, OpenAI got the best models in my opinion
🙏1
Adnan M.
·
·
Claude is more creative, and obeys instructions very well. ChatGPT is good at text manipulation. I use Claude Haiku to create emails that are indistinguishable from human written emails. ChatGPT acted completely stupid and inconsistent.
Dennis H.
·
·
Thank you Adnan! That's the general sense I've been getting too... Out of curiosity, why did you choose the Haiku model specifically? Just cost? or some other bits of performance?
Dennis H.
·
·
fwiw i've been using Opus and i'm wondering if that's overkill
Adnan M.
·
·
Opus is overkill. Haiku is pretty awesome and cheap, too!
🙏1
Elias S.
·
·
A lot comes down to how much data you need to feed it and how good examples you have. If you have just a little bit of text and it doesn't need to be very creative, then a cheap model like haiku or 4o mini can work great. But if you have a ton of info that requires a lot of "IQ" to get the right output for, then the bigger models are better. The new 3.5 Sonnet from Anthropic is incredible.
1