> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tryhamsa.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Choose the model

> Pick the language model behind your agent: providers and models, what each model row tells you, models that need your own key, custom endpoints and temperature.

The model is the language model that reads your prompt and the conversation and decides what the agent says and which tools it calls. This page shows where to choose it, what the model list tells you, and how the choice changes the agent on a call.

## Where to find it

> [tab] Single prompt agent
>
> In the editor's **Configuration** panel, open **LLM configuration**.

> [tab] Flow agent
>
> In the flow builder, click **Global settings** and open **LLM configuration**. One model runs the whole flow: steps can't use a different model, and the [Change agent settings step](https://docs.tryhamsa.com/agents/flow/steps/change-settings.md) doesn't change it.

## Choose a model

**1. Pick the provider**

Choose a **Provider**: OpenAI, Gemini, Groq, or **Custom (OpenAI Compatible)** for your own endpoint.

**2. Pick the model**

Choose a **Model** from the list. Each row shows how fast and how capable the model is (see below).

![LLM configuration in the Configuration panel, with Provider set to OpenAI, Model set to GPT-4.1 Mini and the Temperature slider](https://mintcdn.com/hamsa-452fa79a/IyoYQ-nNrFRmlJOT/images/agents/settings/llm-configuration-light.webp?fit=max\&auto=format\&n=IyoYQ-nNrFRmlJOT\&q=85\&s=8115995321498a3450d84f1a2f3c8197)

**3. Add a key if the model asks for one**

If the row is tagged **Your key**, an **API key** field appears under the picker. See [Models that run on your own key](#models-that-run-on-your-own-key).

**4. Test the agent**

Make a test call and listen for reply speed and how well the agent follows your instructions. See [Test your agent](https://docs.tryhamsa.com/agents/release/test-your-agent.md).

New agents start on **Gemini** with **Gemini 3.5 Flash Lite**.

![The Model list open: each row shows responds in, context and Intelligence values, GPT-6 Luna is tagged Your key and runs on your own API key, and GPT-5 is tagged Deprecated](https://mintcdn.com/hamsa-452fa79a/IyoYQ-nNrFRmlJOT/images/agents/settings/model-list-light.webp?fit=max\&auto=format\&n=IyoYQ-nNrFRmlJOT\&q=85\&s=cdbedba42d82711265dad1bd2a87129a)

## Read the model list

Each model row can show:

| Item | What it tells you |
| - | - |
| **responds in** *n* **s** | Typical time to the first spoken word. Lower is snappier on the phone. |
| **context** | How much text the model can read at once, for example "1M". It limits how long a prompt and conversation can get. |
| **Intelligence** | The Intelligence Index from Artificial Analysis (artificialanalysis.ai). Higher means more capable on their benchmarks. Some models have no score. |
| **Your key** | The model runs on your own provider API key. |
| **Deprecated** | Hamsa no longer includes the model. You can still pick it, but from the date shown it runs on your own API key instead of Hamsa's. |

A model the list no longer offers shows as "*model* (not available any more)". Pick another one.

The models on offer:

| Provider | Models |
| - | - |
| OpenAI | GPT-5, GPT-5 Mini, GPT-5 Nano, GPT-5.4 Mini, GPT-6 Luna, GPT-4.1, GPT-4.1 Mini, GPT-4.1 Nano, GPT-4o, GPT-4o Mini |
| Gemini | Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 3.0 Flash Preview, Gemini 3.1 Flash Lite, Gemini 3.5 Flash Lite |
| Groq | GPT-120-OSS, GPT-20-OSS |
| Custom (OpenAI Compatible) | Any model your endpoint serves |

## Models that run on your own key

Some models run on your own account with the provider. They are tagged **Your key**, and picking one shows an **API key** field.

**1. Open the key picker**

Under the model, click **API key**. It lists your project's [secrets](https://docs.tryhamsa.com/agents/tools/secrets.md).

**2. Pick or create the secret**

Choose the secret that holds your key, or click **New secret** to add one. The key stays a project secret, so the raw key is never stored on the agent.

When a model needs your key and none is set, a banner on the agent says so. Until you add the key or switch to an included model, the agent can't be saved, published or take calls. Some models only move to your key from a date; the banner and the row show it.

## Use your own endpoint

Pick **Custom (OpenAI Compatible)** to use a model you host or a provider Hamsa doesn't list. The endpoint must accept OpenAI-style chat completion requests with tool calls.

| Field | What to enter |
| - | - |
| **Model name** | The model name your endpoint expects, for example `gpt-4o`. |
| **Base URL** | The API base address, for example `https://api.openai.com/v1`. |
| **API key** | A project secret that holds the key for your endpoint. |

All three are required.

## Temperature

> [tab] Single prompt agent
>
> **Temperature** goes from 0 to 1 in steps of 0.1. Lower values make the model more focused and predictable. New agents start at 0.2, and picking another provider or model sets it to 0.1.
>
> GPT-5 and later models fix it at 1.0 and lock the slider ("This model requires temperature to be set to 1.0"). On calls, GPT-5, GPT-5 Mini, GPT-5 Nano and GPT-6 Luna run without a temperature; GPT-5.4 Mini uses the value set here.

> [tab] Flow agent
>
> Flow agents have no temperature control. They run at temperature 0 on every provider, so picking transitions and extracting values stays consistent from call to call.

## What the choice changes on a call

* **Reply speed.** A faster model starts speaking sooner after the caller finishes. Compare **responds in** values, then confirm on a test call.
* **How well it follows instructions.** Long prompts, many tools and strict rules need a more capable model. A short, focused agent can often use a faster, lighter one. Test both on the same calls.
* **How the prompt enhancer writes.** With the enhancer on **Advanced**, Hamsa formats its added sections for the model family (OpenAI, Gemini, others). See [Write a prompt that works on calls](https://docs.tryhamsa.com/agents/build/write-a-prompt.md#what-the-prompt-enhancer-adds).
* **Reply length on Groq.** Groq models stop each reply at 250 tokens. Keep replies short in the prompt, which suits voice anyway.
* **Reasoning.** OpenAI reasoning models (GPT-5 and later) run at the lowest reasoning effort they accept, to keep replies fast.

## Good to know

* Changing the model changes your draft only. Callers get it when you publish.
* The model list comes from Hamsa, so it can change. When an agent uses a deprecated model, a strip under the editor header says so, with the date it moves to your own key and a **Change model** button. Add your key in **LLM configuration** before then, or switch to an included model.

- Some plans only allow a custom endpoint. On those, saving any other model fails with a message about the plan.

## Related

> [card] [Write a prompt that works on calls](https://docs.tryhamsa.com/agents/build/write-a-prompt.md)
>
> What the model reads, and what to write.

> [card] [Secrets](https://docs.tryhamsa.com/agents/tools/secrets.md)
>
> Store provider keys for your models.

> [card] [Call behavior settings](https://docs.tryhamsa.com/agents/settings/call-behavior.md)
>
> Turn taking, silence and call length.

> [card] [Test your agent](https://docs.tryhamsa.com/agents/release/test-your-agent.md)
>
> Hear the difference before you publish.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.