When we started to build LLMAsAService.io there were just a few vendors, and just a few models. That time is gone. Models are constantly changing, their price (generally) dropping, and always a new model to introduce. It was hard for us to maintain our list of models for our own instance, let alone for all of our users.
We’ve changed how we handle models. We have centralized it where all YOU have to do is switch what models you want on and off. Its that simple for OpenAI, Google and Anthropic. Microsoft Azure and Amazon Bedrock models just require a few extra fields – still, much better than hand crafting headers, body and URLs by hand which is what we made you do.
Once logged into the home page, Click on the Admin -> Models link.
New projects will be onboarded with the pre-configured models as a default. So as not to change your current setting for existing customers, you will need to slide the “Use pre-configured models” switch. And if your Agents target specific service groups, you will need to update those to target the new Model Groups.
Our advice going forward is to NEVER use the Advanced Service Configuration tab unless you are targeting private models hosted internally.
The benefits of the new models are –
- Five standard model groups
- Standard
- Premium
- Reasoning/Thinking/Research
- Super Premium
- Super Lite
- We handle pricing and model introduction updates
- We handle setting the list of currently “Suggested” models, which typically give 3-4 models in each group for redundance and reliability
- A button for you to accept our Suggested models, which will be shown to you the first time you go to the model page if you haven’t configured models yet
- A tab where you configure your API keys (optional, you will use our API keys and this will be billed back to your API Credit)
Feedback appreciated. Especially if you notice something out of date.
Team LLMAsAService
Leave a Reply