---
title: Qwen2.5-VL-3B-Instruct | Runware Docs
url: https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct
description: Instruction-tuned vision-language model for image and text understanding
---
# Qwen2.5-VL-3B-Instruct

![Qwen2.5-VL-3B-Instruct](https://assets.runware.ai/covers/alibaba-qwen2-5-vl-3b-instruct.jpg)

Qwen2.5-VL-3B-Instruct is a multimodal model that processes images and text together to perform visual reasoning, captioning, question answering, and structured output tasks. It integrates a vision encoder with an instruction-tuned language backbone to support complex visual understanding and interactive multimodal responses.

- **Status**: live
- **ID**: `alibaba-qwen2-5-vl-3b-instruct`
- Try in Playground
- **Creator**: Alibaba
- **Release Date**: August 24, 2023
- **Zero Data Retention**: Yes
- **Capabilities**: Image to Text, Caption, Checkpoint

## Pricing

$0.003 from, per run · ~7s per request

**Measured runs**

- 1024×1024 input: `$0.003` (~6s)

**4 example generations**

- [caption](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct/examples#example-marine-conservation-field-report-caption): `$0.0013` (~5s)
- [caption](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct/examples#example-fashion-editorial-accessibility-caption): `$0.0019` (~7s)
- [caption](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct/examples#example-urban-flood-response-caption): `$0.0032` (~14s)
- [caption](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct/examples#example-construction-safety-audit-caption): `$0.0032` (~16s)

billed by compute time, so a request costs what it took · times vary, measured end to end · [How pricing works](https://runware.ai/docs/models-api/pricing)

## Request Parameters

**API Options**

Platform-level options for task execution and delivery.

### [taskType](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#request-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `caption`

Identifier for the type of task being performed.

### [taskUUID](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#request-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier for tracking tasks and matching async responses. Must be unique per task.

### [webhookURL](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#request-webhookurl)

- **Type**: `string`
- **Format**: `uri`

Webhook URL that receives JSON responses via HTTP POST when generation tasks complete. For batch requests with multiple results, each completed item triggers a separate webhook call as it becomes available.

**Learn more** (1 resource):

- [Webhooks](https://runware.ai/docs/models-api/webhooks) (platform)

### [deliveryMethod](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#request-deliverymethod)

- **Type**: `string`
- **Default**: `sync`

Determines how the API delivers task results.

**Allowed values**:

- `sync` Returns results directly in the API response. Over HTTP, a task still running after 120 seconds continues as an async task and comes back with status 'processing'. Poll for its results using getResponse.
- `async` Returns an immediate acknowledgment with the task UUID. Poll for results using getResponse.

**Learn more** (1 resource):

- [Task Polling](https://runware.ai/docs/models-api/task-polling) (platform)

### [includeCost](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#request-includecost)

- **Type**: `boolean`

Include task cost in the response.

**Inputs**

Input resources for the task (images, audio, etc). These must be nested inside the \`inputs\` object.

### [image](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#request-inputs-image)

- **Path**: `inputs.image`
- **Type**: `string`
- **Required**: true

Image input (UUID, URL, Data URI, or Base64).

**Core Parameters**

Primary parameters that define the task output.

### [model](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#request-model)

- **Type**: `string`
- **Required**: true
- **Value**: `alibaba-qwen2-5-vl-3b-instruct`

Send `auto` instead of an AIR identifier to have the model router choose the model for this task. The rest of the request is unchanged, every value you send is kept, and only the parameters you leave out are filled in. See [Model Router](https://runware.ai/docs/models-api/model-router).

**Learn more** (3 resources):

- [Text To Image: Model Selection](https://runware.ai/docs/learn/text-to-image#model-selection) (learn)
- [Image Inpainting: Model Specialized Inpainting Models](https://runware.ai/docs/learn/image-inpainting#model-specialized-inpainting-models) (learn)
- [Image Outpainting: Other Critical Parameters](https://runware.ai/docs/learn/image-outpainting#other-critical-parameters) (learn)

### [prompt](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#request-prompt)

- **Type**: `string`

Instructions or questions to guide the image analysis.

## Response Parameters

### [taskType](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#response-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `caption`

Identifier for the type of task this response belongs to.

### [taskUUID](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#response-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier echoed from the original request, used to match async responses to their tasks.

### [text](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#response-text)

- **Type**: `string`
- **Required**: true

Generated caption text describing the input.

### [cost](https://runware.ai/docs/models/alibaba-qwen2-5-vl-3b-instruct#response-cost)

- **Type**: `float`

Task cost in USD. Present when `includeCost` is set to `true` in the request.

## Examples

### caption (caption)

```
A group of marine conservationists is conducting nighttime research on a green sea turtle on a sandy beach. The turtle is illuminated by red lights, likely for identification purposes. One researcher holds a clipboard, while another uses a device, possibly a GPS tracker, attached to the turtle's shell. The team is equipped with various tools and equipment, including a case containing batteries and other devices. The beach appears calm with minimal footprints, suggesting low human activity.
```

```json
{
  "taskType": "caption",
  "taskUUID": "23d0a3f1-71e2-4368-b936-97e8fda943b5",
  "model": "runware:152@1",
  "prompt": "Write an accurate, detailed caption for a professional marine conservation field report. Describe only what is visibly supported by the image, including the main subjects, their actions, the equipment in use, relative positions, beach conditions, lighting, tracks, shoreline, and relevant background activity. Do not invent names, location, date, organization, turtle species, research purpose, measurements, or outcomes. If a detail is uncertain, use cautious visual language rather than guessing. Produce one polished paragraph of 90–130 words in a neutral documentary tone.",
  "inputs": {
    "image": "https://assets.runware.ai/assets/inputs/7fa3243b-ca28-4a35-816b-466e460e2014.jpg"
  }
}
```

---

### caption (caption)

```
In a well-lit, professional makeup station, a model stands poised in a striking blue, ruffled gown as makeup artists meticulously apply finishing touches. The model's vibrant red hair contrasts sharply with her elegant attire. To the right, a rack holds various dresses, including a flowing white gown and a patterned dress, suggesting a diverse wardrobe selection. The floor is marked with white lines, likely indicating designated areas for different tasks.
```

```json
{
  "taskType": "caption",
  "taskUUID": "1c678388-2490-40dd-ab90-794e439df9cf",
  "model": "runware:152@1",
  "prompt": "Write one accurate, detailed accessibility caption for this fashion editorial image. Ground every statement only in clearly visible content. Identify the main subjects, clothing, actions, left-to-right positions, foreground and background relationships, lighting, colors, and notable environmental details. Do not identify people, brands, designers, locations, emotions, or event names unless visibly confirmed. Do not speculate about intent or circumstances. Use clear present-tense prose in 90–130 words.",
  "inputs": {
    "image": "https://assets.runware.ai/assets/inputs/94b27859-ec9d-480c-9adf-a6356e3f85ec.jpg"
  }
}
```

---

### caption (caption)

```
In a rainy urban setting, emergency responders clad in bright orange uniforms are seen assisting a person seated in a small inflatable boat. The boat is being maneuvered by two individuals, one of whom is holding onto a rope connected to the boat. Nearby, a cart filled with stacks of bottled water is positioned on the wet pavement, indicating a possible evacuation or relief effort. In the background, a bus is partially submerged in floodwaters, emphasizing the severity of the situation.
```

```json
{
  "taskType": "caption",
  "taskUUID": "899251f4-adc9-4070-a387-3b00e5d4e890",
  "model": "runware:152@1",
  "prompt": "Write one accurate, detailed, publication-ready caption for this documentary image. Describe the principal subjects, their visible actions, relevant clothing or equipment, clear spatial relationships from foreground to background, and important environmental details. Ground every statement only in what can be directly seen. Do not invent names, dates, locations, organizations, causes, outcomes, dialogue, emotions, or personal relationships. Do not infer identities or circumstances beyond visible evidence. Mention written text only when it is clearly legible. Use neutral journalistic language in a single coherent paragraph.",
  "inputs": {
    "image": "https://assets.runware.ai/assets/inputs/c854e667-6c26-47f0-a6e5-106728cb2d4d.jpg"
  }
}
```

---

### caption (caption)

```
A group of construction workers is gathered under a bridge, engaged in various tasks related to a construction project. They are wearing high-visibility vests and hard hats, indicating adherence to safety protocols. One worker is operating a laser level, which is placed on a tripod, suggesting they are measuring or leveling surfaces. Another worker is holding a rolled-up blueprint, likely reviewing plans or discussing the project's progress. Nearby, there is a crane with its arm extended, possibly lifting materials or equipment.
```

```json
{
  "taskType": "caption",
  "taskUUID": "b8254982-22f6-452d-a72d-212afe27e5c0",
  "model": "runware:152@1",
  "prompt": "Write one accurate, detailed caption for this construction-site photograph. Ground every statement only in visible evidence. Describe the people, their observable actions, clothing and safety equipment, machinery, tools, materials, spatial relationships, site conditions, background, and lighting. Include reliable counts and legible text or markings when visible. Do not infer identities, job titles, project status, hazards, causes, intentions, or events outside the frame. If a detail is unclear, describe it cautiously rather than guessing. Return only the caption.",
  "inputs": {
    "image": "https://assets.runware.ai/assets/inputs/0cd6b11b-1fd1-4656-806b-d429fd6a7b95.jpg"
  }
}
```