Bring your own models
Instead of using one of the pre-deployed Large Language Models (LLM), you may configure custom model that will use provider of your choice and your own API keys. To do this, add your model in the Models screen and specify the provider, model, and API key. Once added, you can select your custom model in the Agent's configuration screen.
The following custom models and providers are currently supported:
| Provider | Models |
|---|---|
| OpenAI | gpt-5.6-terra gpt-5.6-luna gpt-chat-latest gpt-5.5 gpt-5.4 gpt-5.4-mini gpt-5.4-nano gpt-5.2 gpt-5.1 gpt-5 gpt-5-mini gpt-5-nano gpt-realtime (♦) gpt-realtime-mini (♦) gpt-realtime-1.5 (♦) gpt-realtime-2 (♦) gpt-realtime-2.1 (♦) gpt-realtime-2.1-mini (♦) gpt-4.1 gpt-4.1-mini gpt-4.1-nano gpt-4o gpt-4o-mini gpt-4o-realtime (♦) gpt-4o-mini-realtime (♦) gpt-oss-120b gpt-oss-20b |
| Azure OpenAI | gpt-chat-latest gpt-5.5 gpt-5.4 gpt-5.4-mini gpt-5.4-nano gpt-5.2 gpt-5.1 gpt-5 gpt-5-mini gpt-5-nano gpt-realtime (♦) gpt-realtime-mini (♦) gpt-realtime-1.5 (♦) gpt-realtime-2 (♦) gpt-realtime-2.1 (♦) gpt-realtime-2.1-mini (♦) gpt-4.1 gpt-4.1-mini gpt-4.1-nano gpt-4o gpt-4o-mini gpt-4o-realtime (♦) gpt-4o-mini-realtime (♦) gpt-oss-120b gpt-oss-20b |
| Google (API) | gemini-3.8-flash gemini-3.8-live (♦) gemini-3.8-live-extended-thinking (♦) gemini-3.6-flash gemini-3.5-flash gemini-3.5-flash-lite gemini-3.1-flash-lite gemini-3.1-flash-live (♦) gemini-3-flash gemini-2.5-flash gemini-2.5-flash-lite gemini-2.5-flash-native-audio (♦) gemini-live-2.5-flash (♦) gemma-4-31b-it gemma-4-26b-a4b-it |
| Google (Enterprise) | gemini-3.8-flash gemini-3.8-live (♦) gemini-3.6-flash gemini-3.5-flash gemini-3.5-flash-lite gemini-3.1-flash-lite gemini-3-flash gemini-2.5-flash gemini-2.5-flash-lite gemini-2.5-flash-native-audio (♦) gemma-4-31b-it gemma-4-26b-a4b-it |
| Anthropic | claude-sonnet-5 claude-sonnet-4-5 claude-haiku-4-5 |
| Groq | llama-3.1-8b-instant llama-3.3-70b-versatile openai/gpt-oss-120b openai/gpt-oss-20b qwen/qwen3.6-27b |
| Cerebras | gpt-oss-120b gemma-4-31b zai-glm-4.7 |
| Amazon | nova-micro nova-lite nova-pro nova-sonic (♦) nova-2-lite nova-2-sonic (♦) |
| Mistral | mistral-small mistral-medium mistral-large ministral-8b ministral-14b |
| xAI | grok-4.5 grok-4.3 grok-4-1-fast-reasoning grok-4-1-fast-non-reasoning grok-voice (♦) |
| Custom | Any model served over the OpenAI Chat Completions API — see Custom models |
Models marked with (♦) are realtime models. For details, see Speech-to-speech models
Some models support more than one provider API — for example, OpenAI's Chat Completions or Responses API, and Gemini's OpenAI-compatible or native Gemini API. The platform automatically selects the optimal API for each model, but you can override this choice. See Selecting the model API for details.
Custom models
The "Custom" provider lets you bring any model that is served over the OpenAI Chat Completions API — a model you host yourself (for example, with vLLM, Ollama or llama.cpp), an LLM gateway in front of your own models, or a provider that isn't listed above but offers an OpenAI-compatible endpoint.
The platform has no built-in knowledge of such a model, so instead of picking it from a list you type its name, enter the endpoint's base URL, and tell the platform what the model supports: its context length, how it calls tools, whether it streams, whether it supports structured output, and whether it writes its reasoning into the reply. For a description of each setting, see Custom provider.
A custom model:
- Is a text model: speech-to-speech (realtime) models can't be added through the "Custom" provider.
- Always runs over Chat Completions — Selecting the model API doesn't apply to it, even when it serves a
gpt-*or Gemini model.