Hi everyone,
I’ve been an active Groq user since around August, both in the forums and as a heavy API user for my platform, Lernex, an AI-powered learning platform where fast inference is extremely important.
When I first started using Groq, I was genuinely impressed. The pricing was competitive, inference speeds were fantastic, and Groq hosted the exact models I needed at the time (especially GPT-OSS-20B and GPT-OSS-120B). It felt like Groq had a very promising future as a serious inference provider.
However, after about 7–8 months of using the platform, I’ve started to feel disappointed with the lack of progress in the hosted model ecosystem.
Since I began using Groq, I haven’t seen a single new model added to the production stack. During that same time, the open-source ecosystem has moved extremely quickly. There have been dozens of new models released that many developers would likely want access to.
I’ve also noticed that the community forums contain hundreds of requests for new hosted models, and Groq staff often respond to these posts. But from what I can tell, none of those requests have actually resulted in new models being deployed.
For developers building real products on top of Groq, this is concerning. Fast inference alone isn’t enough if the available models become outdated relative to what’s available elsewhere.
I’m still hopeful about Groq’s long-term potential, and I genuinely want the platform to succeed. But it would be great to see:
-
More regular additions of new open-source models
-
Clearer communication about the roadmap for hosted models
-
Some indication that community requests are translating into real deployments
Right now it feels like the community is providing a lot of feedback and enthusiasm, but we’re not seeing much movement on the platform side.
Would love to hear if there are plans to expand the hosted model lineup in the near future.
