Groq is impressive! Here’s a brief review: ⤵️
• Decidedly fast 🏎️: it doesn’t take long to “think”, it just responds!!
• Quality of Mistral 7b and Llama 2-70B is decent 👍: not always GPT4-level but still given the speed it’s a great brainstorming-buddy/co-author
• Code-generation quality is ok 🔢: both Mistral and Llama do a decent job (even with low-context prompt such as “a value added bar chart a la McKinsey in the style of economist”
• Open source 👐: it’s great to know I could run both models locally (or on-prem) if I wanted to.
• Multi-modal: missing. ChatGPT still the king.
In conclusion: we'll be using Groq’s app and API more as its speed is just amazing for a lot of day to day needs. Excited to see more.
P.S.: here’s a telling comparison. Tell us in the comments which jingle about JWT you like the most!!
Негізгі бет Groq vs ChatGPT
Пікірлер