[Integrations](https://spiral.ai/integrations) › [Claude Code](https://spiral.ai/integrations/claude-code) › [Qwen3 VL 235B A22B Instruct](https://spiral.ai/integrations/claude-code/qwen3-vl-235b-a22b-instruct)

# How to Use Qwen3 VL 235B A22B Instruct in Claude Code

Qwen3 VL 235B A22B Instruct reads documents, screenshots, and video for extraction and visual coding. Unlike Qwen3 VL 235B A22B Thinking, it uses the non-thinking format; Thinking is stronger on complex visual reasoning.

|  |  |
| --- | --- |
| Model | Qwen3 VL 235B A22B Instruct |
| Available through | OpenRouter |
| Category | LLM |
| Price | Up to $0.010127 per run at the default inputs. $0.273 per 1M input tokens, $2.47 per 1M output tokens. |
| Returns | text, json |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with Qwen3 VL 235B A22B Instruct in Claude Code

- Use Qwen3 VL 235B A22B Instruct to interpret the reference screenshot in this project. Implement the layout using our existing components, preserve the visible copy, and run the relevant checks.
- Use Qwen3 VL 235B A22B Instruct to extract headings and reading order from the scanned sample pages in this repository. Save an ordered outline with page references and flag uncertain text.
- Use Qwen3 VL 235B A22B Instruct to read the receipt images in our test fixtures. Compare the extracted merchant, date, and total with our expected records and report discrepancies.
- Use Qwen3 VL 235B A22B Instruct to inspect the app screenshot in this project. Locate the search control, describe nearby labels, and propose a browser test using the actual DOM.
- Use Qwen3 VL 235B A22B Instruct to draft a frontend from the supplied design image. Follow this project's framework and styles, then list visual details that still need manual comparison.

## How to set up Qwen3 VL 235B A22B Instruct in Claude Code

Pick the way that suits you. Each takes about two minutes.

### Use Spiral's MCP server

1. **Add the hosted server**: This adds Spiral to the current project. Add `--scope user` to use it in every project.

```bash
claude mcp add --transport http spiral https://api.spiral.ai/mcp
```

2. **Sign in**: Start `claude`, type `/mcp`, choose spiral and Authenticate. Sign in to Spiral in the browser and click Authorize. No API key.

### Use the Spiral skill

1. **Send Claude Code this message**: Your agent reads Spiral's skill and sets itself up.

```
set up https://spiral.ai/SKILL.md
```

2. **Sign in or paste a key**: Approve Spiral in the browser, or give the agent a key from the console.

### Use the Spiral CLI

1. **Install and sign in**: Installs the Spiral CLI, signs you in and adds its skill to your coding agents.

```bash
npx -y spiral-cli@latest init
```

### Call the Spiral API

1. **Create an API key**: In the console, under API keys.

2. **Run Qwen3 VL 235B A22B Instruct**: One request; the result comes back by polling or a webhook.

```bash
curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "openrouter",
    "model": "qwen/qwen3-vl-235b-a22b-instruct",
    "input": {
      "messages": [
        {
          "role": "system"
        }
      ]
    }
  }'
```

Then paste one of the prompts above: Claude Code finds Qwen3 VL 235B A22B Instruct on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/claude-code

## Qwen3 VL 235B A22B Instruct tools and prices

- [Qwen: Qwen3 VL 235B A22B Instruct](https://spiral.ai/tools/openrouter/qwen/qwen3-vl-235b-a22b-instruct) (OpenRouter): Up to $0.010127 per run at the default inputs. $0.273 per 1M input tokens, $2.47 per 1M output tokens.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| seed |  | — |  |
| stop |  | — |  |
| min_p | 0 to 1 | — |  |
| tools |  | — |  |
| top_k | 0 to … | — |  |
| top_p | 0 to 1 | — |  |

## What can you do with Qwen3 VL 235B A22B Instruct in Claude Code?

Turn visual references into code drafts and document images into extraction tasks. A coding agent can use Qwen3 VL 235B A22B Instruct alongside project files to implement the result and check it.

### Extract invoice details

Turn invoice images into records containing the vendor, invoice number, date, and total. Request a supporting text excerpt for each value and leave missing values empty.

### Transcribe multilingual signs

Read photographed signs or labels in their original writing. Keep line breaks and mark uncertain characters for review.

### Draft a page from a screenshot

Use a screenshot as the reference for a frontend implementation. Specify the framework and request the visible text, layout, and controls in the code draft.

### Locate visual elements

Find a named object or interface element in an image and describe its location. Request nearby visual landmarks so a reviewer can check the identification.

### Summarize a demonstration

Turn the visible steps in a demonstration video into a written procedure. Ask for observed actions and a separate list of steps that need clarification.

### Parse document layouts

Convert scanned document pages into an ordered outline with headings, paragraphs, and extracted text. Preserve the page references for checking against the originals.

## Overview

- [OpenRouter](https://spiral.ai/providers/openrouter)
- [All Claude Code guides](https://spiral.ai/integrations/claude-code)

## Qwen3 VL 235B A22B Instruct in other agents

- [Amp](https://spiral.ai/integrations/amp/qwen3-vl-235b-a22b-instruct)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/qwen3-vl-235b-a22b-instruct)
- [Claude](https://spiral.ai/integrations/claude-desktop/qwen3-vl-235b-a22b-instruct)
- [Cline](https://spiral.ai/integrations/cline/qwen3-vl-235b-a22b-instruct)
- [Codex](https://spiral.ai/integrations/codex/qwen3-vl-235b-a22b-instruct)
- [Cursor](https://spiral.ai/integrations/cursor/qwen3-vl-235b-a22b-instruct)
- [Devin](https://spiral.ai/integrations/devin/qwen3-vl-235b-a22b-instruct)
- [dots](https://spiral.ai/integrations/dots/qwen3-vl-235b-a22b-instruct)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/qwen3-vl-235b-a22b-instruct)
- [Gemini](https://spiral.ai/integrations/gemini/qwen3-vl-235b-a22b-instruct)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/qwen3-vl-235b-a22b-instruct)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/qwen3-vl-235b-a22b-instruct)
- [goose](https://spiral.ai/integrations/goose/qwen3-vl-235b-a22b-instruct)
- [Grok](https://spiral.ai/integrations/grok/qwen3-vl-235b-a22b-instruct)
- [Grok Build](https://spiral.ai/integrations/grok-build/qwen3-vl-235b-a22b-instruct)
- [Hermes](https://spiral.ai/integrations/hermes/qwen3-vl-235b-a22b-instruct)
- [Instinct](https://spiral.ai/integrations/instinct/qwen3-vl-235b-a22b-instruct)
- [Manus](https://spiral.ai/integrations/manus/qwen3-vl-235b-a22b-instruct)
- [Muse](https://spiral.ai/integrations/muse/qwen3-vl-235b-a22b-instruct)
- [Muse Code](https://spiral.ai/integrations/muse-code/qwen3-vl-235b-a22b-instruct)
- [n8n](https://spiral.ai/integrations/n8n/qwen3-vl-235b-a22b-instruct)
- [Notion](https://spiral.ai/integrations/notion/qwen3-vl-235b-a22b-instruct)
- [OpenClaw](https://spiral.ai/integrations/openclaw/qwen3-vl-235b-a22b-instruct)
- [opencode](https://spiral.ai/integrations/opencode/qwen3-vl-235b-a22b-instruct)
- [Perplexity](https://spiral.ai/integrations/perplexity/qwen3-vl-235b-a22b-instruct)
- [Raycast](https://spiral.ai/integrations/raycast/qwen3-vl-235b-a22b-instruct)
- [VS Code](https://spiral.ai/integrations/vscode/qwen3-vl-235b-a22b-instruct)
- [Windsurf](https://spiral.ai/integrations/windsurf/qwen3-vl-235b-a22b-instruct)
- [Zapier](https://spiral.ai/integrations/zapier/qwen3-vl-235b-a22b-instruct)

## More Claude Code guides

- [Qwen3 Next 80B A3B Instruct](https://spiral.ai/integrations/claude-code/qwen3-next-80b-a3b-instruct)
- [Qwen3 VL 30B A3B Instruct](https://spiral.ai/integrations/claude-code/qwen3-vl-30b-a3b-instruct)
- [Qwen3 Coder Next](https://spiral.ai/integrations/claude-code/qwen3-coder-next)
- [Qwen3-VL 8B Instruct](https://spiral.ai/integrations/claude-code/qwen3-vl-8b-instruct)
- [Qwen3 Coder Flash](https://spiral.ai/integrations/claude-code/qwen3-coder-flash)
- [Qwen3.5 397B A17B](https://spiral.ai/integrations/claude-code/qwen3-5-397b-a17b)

## Questions about Qwen3 VL 235B A22B Instruct in Claude Code

**Can Claude Code use Qwen3 VL 235B A22B Instruct?**

Yes. Add Spiral to Claude Code as an MCP server, and Claude Code can find and run Qwen3 VL 235B A22B Instruct like any built-in tool.

**How much does Qwen3 VL 235B A22B Instruct cost in Claude Code?**

Up to $0.010127 per run at the default inputs. $0.273 per 1M input tokens, $2.47 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Claude Code plan is separate.

**Do I need an OpenRouter account to use Qwen3 VL 235B A22B Instruct in Claude Code?**

No. Spiral runs Qwen3 VL 235B A22B Instruct on its own OpenRouter account. You sign in to Spiral once and never paste a key.

**How long does it take to set up Qwen3 VL 235B A22B Instruct in Claude Code?**

About two minutes: add the hosted server, then sign in.

**Does Claude Code wait while Qwen3 VL 235B A22B Instruct runs?**

Qwen3 VL 235B A22B Instruct runs in the background. Claude Code starts the run, checks on it and gives you the text, json when it's ready.

**What is Qwen3 VL 235B A22B Instruct useful for?**

Use it to extract information from document images, read photographed text, interpret visual content, and draft code from screenshots. Ask for a concrete deliverable, such as an extraction table or an implementation draft.

**How does Instruct compare with Thinking?**

Instruct uses the non-thinking format. Qwen reports stronger complex visual reasoning from Thinking, so Instruct is a better starting point for direct extraction tasks than for difficult visual proofs.

**Can it read text in multiple languages?**

Qwen documents OCR support for 32 languages. For transcription, request the original writing and mark unreadable spans explicitly.

**How much context and output does Qwen document?**

Qwen documents a native 256K-token context window. Its Instruct evaluations use a maximum output length of 32,768 tokens; treat that number as the published evaluation setting.

**Can it turn a screenshot into code?**

Yes. Qwen documents code generation from images and videos; specify the target language, visible content to preserve, and how you will check the result.

