Recommended Language Models for Concierge
Here is a list of Large Language Models (LLM) that we recommend to use with Concierge Service.
| API Type | Language Model | Notes |
|---|---|---|
OpenAI | GPT-5.5, GPT-5.4 | All models from GPT-5 family *except -mini, -nano, -pro, and -codex |
gemini-2.5-pro gemini-3.5-flash** |
** -flash supported starting in v7.2.1; may need a patch if running into 400 error | |
| Azure OpenAI | gpt-5.5, gpt-5.4 | -chat versions should work as well, but they change very quickly and are not well documented. |
| Anthropic | claude-opus-4-7, claude-sonnet-4-6 |
* We do not recommend using -mini, -nano, -pro or -codex models. The quality of reasoning and generation for -mini and -nano versions is significantly lower than the full model versions. The -pro version is rather pricey and frankly overkill for this use case. The -codex version is not supported at all due to different APIs.