I am currently configuring the LLM integration in my Teable panel, but I’ve run into a couple of questions regarding model selection.
Right now, it seems I am only able to use the GPT 5.5 model in my panel. I would really like to switch to Claude Opus 4.7, but I couldn’t find an option to change it.
I would appreciate some clarity on the following:
How can I switch to the Claude Opus 4.7 model? Is this currently supported, and if so, where can I configure it?
Are there other LLMs available for users to choose from, or is the panel currently restricted to specific models?
A quick suggestion for the Teable team: While it makes perfect sense for Teable to recommend certain models by default in the backend/admin panel, please don’t restrict users from choosing their preferred LLMs. Having the flexibility to select and swap models based on our personal preference or specific workflows would make Teable significantly more powerful.
Hi @OtakuSay, We recently updated the model selection because GPT-5.5 currently offers stronger overall intelligence, significant cost advantages, and a more user-friendly experience. We made this change with the goal of letting users use Teable freely without worrying about credits.
That said, thank you for your suggestion and detailed feedback. We understand that you’d like to have more model options. We’ll keep this feedback in mind and continue to evaluate it, but we don’t plan to change the current model selection setup in the short term. I wanted to set the right expectation upfront so it doesn’t affect how you plan to use the product.
Please rest assured that we’ll continue optimizing Teable in the most user-friendly and high-quality way possible, and we’ll always work to make Teable the best experience in its category.
ChatGPT is slow and takes forever to execute requests. Also it’s not as good as Claude in App building or understanding complex queries. We should have the option to select LLM. I don’t like these new changes.
Hi Leo,
GPT is much more cost‑effective than Claude, but when I try to build an application it is very slow and the results are worse than Claude. In my opinion, even MiniMax 2.7 works smarter and faster with low credit cost. It will be greate to have the option to choose the model that we want, but also I understand your reasoning.
We’ve identified one issue that may affect some cases: GPT is very strict when handling secrets, such as API keys, SMTP credentials, and other sensitive values. In some situations, even if users provide these credentials, Cuppy may avoid using them directly, which can block the task from moving forward.
We’re working on a dedicated secrets / environment variables management feature to improve this. Once it’s available, you should be able to provide credentials more safely in Teable, and Cuppy should be able to use them when helping you in conversations.
This feature is already in development and should be available soon.
If your case is not related to API keys, SMTP credentials, or other secrets, please share more details here. We’ll review those cases separately.
Hi Leo, this would be amazing and this will remove a lot of headache.
Nevertheless, being able to select other models (Claude, MiniMax) would be nice as well. As @jakub.stec mentioned, MiniMax 2.7 was super fast, cost effective and for what I needed, better than ChatGPT 5.5 (or even 5.4 mini)
I understand why MiniMax feels attractive. I’ve also tried MiniMax 2.7 Highspeed myself. It is indeed very fast, but in my own test, it failed to complete the task and cost me around 30 minutes of extra cleanup and rework. In my case, the faster response speed did not save time overall.
Hi @Alex ,
I understand why MiniMax feels attractive. I’ve also tried MiniMax 2.7 Highspeed myself. It is indeed very fast, but in my own test, it failed to complete the task and cost me around 30 minutes of extra cleanup and rework. In my case, the faster response speed did not save time overall.