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

# Models

> Pick the right model for each thread.

Every thread has a model picker. Switch anytime, even mid-conversation.

## Available models

| Model                     | Provider    | Vision | Reasoning | Best for                           |
| ------------------------- | ----------- | ------ | --------- | ---------------------------------- |
| **Fable 5**               | Anthropic   | Yes    | High      | Complex, multi-step automations    |
| **Opus 5**                | Anthropic   | Yes    | High      | Deep reasoning and agentic coding  |
| **Sonnet 5**              | Anthropic   | Yes    | Medium    | Strong all-rounder                 |
| **Kimi K3**               | Moonshot AI | Yes    | Medium    | Long-context reasoning with vision |
| **Grok 4.5**              | xAI         | Yes    | Medium    | Frontier coding and STEM reasoning |
| **GPT-5.5** (default)     | OpenAI      | Yes    | Low       | Fast general tasks                 |
| **Gemini 3.1 Flash Lite** | Google      | Yes    | None      | Quick lookups and light browsing   |
| **GLM 5.2**               | Z.ai        | No     | Medium    | Text-heavy work at low cost        |
| **DeepSeek V4 Pro**       | DeepSeek    | No     | Medium    | Reasoning on text-only tasks       |

## Vision and browsing

Models with vision see screenshots of the live browser, which helps on visually complex pages. Text-only models (GLM 5.2, DeepSeek V4 Pro) still browse using the page structure; Quaso strips screenshots from the context automatically for them.