# Model Library

Fetch Hive supports multiple model providers across prompts, workflow AI steps, and agent settings.

The list below reflects the models shown in the shared model picker used across the product.

Models that support extended reasoning display a **Reasoning** badge in the model picker. Depending on the provider, reasoning surfaces as one of three UI controls in the parameters panel:

* **Max Thinking Tokens** — Anthropic Claude and Google Gemini thinking models. The slider controls how many internal reasoning tokens the model uses before generating a response.
* **Reasoning Effort** — OpenAI and xAI reasoning models. A dropdown (Low / Medium / High) that trades compute for response quality.
* **Reasoning toggle** — Together AI reasoning models. An Enabled/Disabled dropdown that switches reasoning on or off for each prompt or workflow step.

## OpenAI

| Model        | Vision | Reasoning    |
| ------------ | ------ | ------------ |
| GPT 5        | Yes    | Yes (effort) |
| GPT 5 Mini   | Yes    | Yes (effort) |
| GPT 5 Nano   | Yes    | Yes (effort) |
| GPT 4.1      | Yes    | No           |
| GPT 4.1 Mini | Yes    | No           |
| GPT 4.1 Nano | Yes    | No           |
| o4 Mini      | No     | Yes (effort) |
| o3           | No     | Yes (effort) |
| o3 Mini      | No     | Yes (effort) |
| o1           | No     | Yes (effort) |
| o1 Pro       | No     | Yes (effort) |

## Anthropic

| Model             | Vision | Reasoning             |
| ----------------- | ------ | --------------------- |
| Claude Sonnet 4.6 | Yes    | Yes (thinking tokens) |
| Claude Sonnet 4.5 | Yes    | Yes (thinking tokens) |
| Claude Opus 4.6   | Yes    | Yes (thinking tokens) |
| Claude Opus 4.5   | Yes    | Yes (thinking tokens) |
| Claude Haiku 4.5  | Yes    | No                    |

## Google

| Model                 | Vision | Reasoning             |
| --------------------- | ------ | --------------------- |
| Gemini 2.5 Pro        | Yes    | Yes (thinking tokens) |
| Gemini 2.5 Flash      | Yes    | Yes (thinking tokens) |
| Gemini 2.5 Flash Lite | Yes    | No                    |
| Gemini 2.0 Flash      | Yes    | No                    |
| Gemini 2.0 Flash Lite | Yes    | No                    |

## xAI

| Model    | Vision | Reasoning    |
| -------- | ------ | ------------ |
| Grok 4.2 | Yes    | Yes (effort) |
| Grok 4.1 | Yes    | Yes (effort) |

## Perplexity

| Model               | Vision | Reasoning |
| ------------------- | ------ | --------- |
| Sonar Pro           | No     | No        |
| Sonar               | No     | No        |
| Sonar Reasoning Pro | No     | No        |
| Sonar Reasoning     | No     | No        |
| Sonar Deep Research | No     | Yes       |

## TogetherAI-hosted models

Fetch Hive provides access to a large and frequently updated set of models hosted by [Together AI](https://www.together.ai/), including models from DeepSeek, Qwen, Kimi (Moonshot), Mistral, MiniMax, Meta Llama, and others. Some Together AI models support reasoning — for these, a **Reasoning** (enabled/disabled) toggle appears in the parameters panel.

Because Together AI adds and removes models on a rolling basis, the list is **refreshed automatically every day** from the Together AI API (`GET https://api.together.ai/v1/models`). Only models with active pricing are included; any model that Together AI withdraws or de-prices disappears from the selector on the next daily refresh.

To retrieve the current live list call:

```
GET /v1/models
```

Filter the response for `provider` values such as `deepseek`, `qwen`, `kimi`, `mistral`, `minimax`, or `llama` to see the Together-hosted subset.

## API Access

You can retrieve the current list of active models programmatically via the API:

**`GET /v1/models`** — Returns all active models with provider, context limit, and capability flags. Requires Bearer authentication. See [API Reference: Models](/api-reference/models.md).


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.fetchhive.com/resources/model-library.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.