Description #

Provides deep reasoning capabilities for large models in multimodal scenarios
Uses models with deep reasoning capabilities to analyze and understand images, videos, or text, and returns structured text output. The operator automatically constructs a message structure that complies with multimodal model specifications. Users only need to provide image, video, or text data as per the convention to complete inference.

Key features #

• Deep reasoning mechanism: The model automatically decomposes questions and performs logical reasoning before answering, generating a reasoning chain (reasoning_content)
• Multimodal scenario support: Supports image, video, and text input simultaneously, and automatically assembles multimodal messages
• Input simplification mechanism: Supports multiple data sources such as local files, HTTP/HTTPS URLs, and TOS/S3 object storage. Visual understanding capabilities can be achieved through simple configuration.
• Flexible reasoning mode: Supports controlling the deep reasoning mode (enabled / disabled / auto) via the thinking_type parameter, allowing flexible trade-offs between answer quality and performance

Input and output specifications #

• Input format:
  • Images (images): Supports string or list format
    • Supported data source types:
      • url: Network resource address, supports http/https/tos/s3. TOS/S3 will automatically generate a pre-signed URL
      • base64: Base64-encoded data
      • binary: Binary data (internally converted to Base64 encoding)
  • Videos (videos): Supports string or list format, with data source types consistent with images
  • Texts (texts): Supports string or list format, used as user prompts

(Mixing string and list types in the same field is not allowed)

• Output format:
  • Default mode: Returns a struct type containing the following fields:
    • llm_result: The model's final answer content
    • reasoning_content: The model's reasoning chain / reasoning process
  • When the environment variable LAS_LLM_FINISH_REASON_CHECK=true is set, the struct will additionally include:
    • finish_reason: The reason for the model result ending, with possible values including:
      • stop: Normal termination
      • length: Exceeded token limit
      • content_filter: Triggered content filtering policy
  • When the environment variable LAS_LLM_BOTS_REFERENCES=true is set, the struct will additionally include:
    • references: Retrieval results or external knowledge information referenced by the model, facilitating downstream transmission and parsing

Enhanced model capabilities #

• Chain-of-thought visualization: The reasoning_content field outputs the model's reasoning process, making it easier to observe and review.
• Result reliability control: The finish_reason field identifies abnormal termination scenarios, such as length truncation or content filtering.
• Multimodal understanding: Supports mixed input parsing for images, videos, and text, suitable for complex multimodal scenarios.

Chat #

Interface description #

Use the Chat interface to invoke the doubao-seed-2.0 model for multimodal understanding.
Supported models and versions:

• doubao-seed-2-0-pro-260215
• doubao-seed-2-0-lite-260215
• doubao-seed-2-0-mini-260215

Request parameters #

Response parameters #

Examples #

Request example

curl --location "https://operator.las.ap-southeast-1.volces.com/api/v1/chat/completions" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer $$LAS_API_KEY" \
--data '{
    "messages": [
        {
            "role": "user",
            "content": [
                {
                    "type": "video_url",
                    "video_url": {
                        "url": "https://las-ai-cn-beijing.tos-cn-beijing.volces.com/*****/sample.mp4"
                    }
                },
                {
                    "type": "text",
                    "text": "视频里面有什么"
                }
            ]
        }
    ],
    "model": "doubao-seed-2-0-pro-260215"
}'

Response example

{
  "model": "doubao-seed-2-0-pro-260215",
  "usage": {
    "prompt_tokens": 3877,
    "completion_tokens": 493,
    "total_tokens": 4370,
    "prompt_tokens_details": {
      "cached_tokens": 0,
      "provisioned_tokens": null
    },
    "completion_tokens_details": {
      "reasoning_tokens": 354,
      "provisioned_tokens": null
    }
  },
  "created": 1772692173,
  "id": "0217726921599114f28f9638b6effb5ab036b3eddba1d5642ea3d",
  "choices": [
    {
      "finish_reason": "stop",
      "index": 0,
      "logprobs": null,
      "message": {
        "content": "视频展现的是海边惬意的户外场景：\n1. 核心内容：一位戴着浅米色宽檐帽、黑色墨镜的女士，穿着白色上衣+印花彩色半身裙，靠在白色建筑的墙角，坐在折叠椅上手风琴演奏，演奏时还伴随哼唱，状态十分放松。她的脚边放着一瓶饮用水，背景是蓝色的海面和远处的低矮岸线。\n2. 画面变化：过程中镜头逐渐向女士推近，能越来越清晰看到她演奏的细节和神态；后半段还有穿绿短袖、浅色长裤的路人从她旁边经过。\n整体氛围闲适，充满海滨区域的松弛感。",
        "reasoning_content": "\n用户现在需要描述视频里的内容对吧，首先先梳理场景：首先是户外，靠海的地方，有个白色建筑的墙角对吧？然后主角是个女士，戴着浅色宽檐帽，白上衣，花裙子，墨镜，坐在折叠椅上，正在拉手风琴，旁边地上放着瓶装水，背景是海和远处的岸。然后过程里她一边拉一边还在唱，镜头慢慢往她那边推近对吧？然后后面有路人经过，比如穿绿短袖浅色裤子的人从旁边走过，还有其他路人？哦对，整体氛围很悠闲，是海边的街头演奏那种感觉对吧？\n等下整理清楚：\n首先整体场景：这是在海边的户外，靠近一处白色建筑的墙角，地面是石质地，背景能看到蓝色的海面和远处的低矮陆地，天气晴朗。\n然后核心人物：一位女士坐在黑色折叠椅上，她戴着浅米色宽檐帽、黑色墨镜，穿白色上衣、印着花卉图案的彩色半身裙，脚上是黑色凉鞋，正演奏着手风琴，演奏的过程中她还在跟着哼唱，状态很放松悠闲。她脚边放着一瓶透明的饮用水，旁边还有琴盒类的黑色箱子。\n然后镜头变化：视频过程中镜头逐渐向这位演奏的女士推近，到后半段能清晰看到她演奏的细节和神态。\n最后还有小细节：视频后半段有路人从她旁边的建筑旁走过，比如一位穿薄荷绿短袖、浅卡其色长裤的女士从旁边经过。\n整个氛围很闲适，充满海边小镇的惬意感。\n对，这样说清楚就可以了。",
        "role": "assistant"
      }
    }
  ],
  "object": "chat.completion"
}

Responses #

Interface description #

Use the Responses interface to invoke the doubao-seed-2.0 model for multimodal understanding. Refer to the interface definition in Create a model response.
Supported models and versions:

• doubao-seed-2-0-pro-260215
• doubao-seed-2-0-lite-260215
• doubao-seed-2-0-mini-260215

Examples #

Request example

curl --location "https://operator.las.ap-southeast-1.volces.com/api/v1/responses" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer $LAS_API_KEY" \
--data '{
    "input": [
        {
            "role": "user",
            "content": [
                {
                    "type": "input_video",
                    "video_url": "https://las-ai-cn-beijing.tos-cn-beijing.volces.com/*****/sample.mp4"
                },
                {
                    "type": "input_text",
                    "text": "视频里面有什么"
                }
            ]
        }
    ],
    "model": "doubao-seed-2-0-pro-260215"
}'

Response example

{
    "model": "doubao-seed-2-0-pro-260215",
    "usage": {
        "prompt_tokens": 3877,
        "completion_tokens": 493,
        "total_tokens": 4370,
        "prompt_tokens_details": {
            "cached_tokens": 0,
            "provisioned_tokens": null
        },
        "completion_tokens_details": {
            "reasoning_tokens": 354,
            "provisioned_tokens": null
        }
    },
    "created": 1772692173,
    "id": "0217726921599114f28f9638b6effb5ab036b3eddba1d5642ea3d",
    "choices": [
        {
            "finish_reason": "stop",
            "index": 0,
            "logprobs": null,
            "message": {
                "content": "视频展现的是海边惬意的户外场景：\n1. 核心内容：一位戴着浅米色宽檐帽、黑色墨镜的女士，穿着白色上衣+印花彩色半身裙，靠在白色建筑的墙角，坐在折叠椅上手风琴演奏，演奏时还伴随哼唱，状态十分放松。她的脚边放着一瓶饮用水，背景是蓝色的海面和远处的低矮岸线。\n2. 画面变化：过程中镜头逐渐向女士推近，能越来越清晰看到她演奏的细节和神态；后半段还有穿绿短袖、浅色长裤的路人从她旁边经过。\n整体氛围闲适，充满海滨区域的松弛感。",
                "reasoning_content": "\n用户现在需要描述视频里的内容对吧，首先先梳理场景：首先是户外，靠海的地方，有个白色建筑的墙角对吧？然后主角是个女士，戴着浅色宽檐帽，白上衣，花裙子，墨镜，坐在折叠椅上，正在拉手风琴，旁边地上放着瓶装水，背景是海和远处的岸。然后过程里她一边拉一边还在唱，镜头慢慢往她那边推近对吧？然后后面有路人经过，比如穿绿短袖浅色裤子的人从旁边走过，还有其他路人？哦对，整体氛围很悠闲，是海边的街头演奏那种感觉对吧？\n等下整理清楚：\n首先整体场景：这是在海边的户外，靠近一处白色建筑的墙角，地面是石质地，背景能看到蓝色的海面和远处的低矮陆地，天气晴朗。\n然后核心人物：一位女士坐在黑色折叠椅上，她戴着浅米色宽檐帽、黑色墨镜，穿白色上衣、印着花卉图案的彩色半身裙，脚上是黑色凉鞋，正演奏着手风琴，演奏的过程中她还在跟着哼唱，状态很放松悠闲。她脚边放着一瓶透明的饮用水，旁边还有琴盒类的黑色箱子。\n然后镜头变化：视频过程中镜头逐渐向这位演奏的女士推近，到后半段能清晰看到她演奏的细节和神态。\n最后还有小细节：视频后半段有路人从她旁边的建筑旁走过，比如一位穿薄荷绿短袖、浅卡其色长裤的女士从旁边经过。\n整个氛围很闲适，充满海边小镇的惬意感。\n对，这样说清楚就可以了。",
                "role": "assistant"
            }
        }
    ],
    "object": "chat.completion"
}

Error codes #