4 ms·
How is openrouter working out?
by mmaunder 2y ago
How is openrouter working out?
- bobcoi 2y agoreally nice. Just one API you get access to 100+ models If I want to change models I just change one line of code. Pricing is pretty fair as well they just charge a % on top
- mmaunder 2y agoDo you think you'll switch to running local models at some point e.g. LambdaLabs/Runpod running Llama to reduce costs? Just thinking about as you scale and wanting to reduce cost per token. (I'm not in the GPU hosting biz - just curious because I think we're all thinking about this)