Qwen3.8 27B

Summary: Qwen3.8-27B is a compact dense multimodal model from Alibaba Cloud's Qwen team and the most capable generation in the Qwen open-model family to date. Built on the architectural foundation of Qwen3.5, it pairs a 27B parameter language model with an integrated vision encoder and natively supports a 262K-token context window. The model delivers substantial gains in coding, professional work, research, and long-horizon agentic tasks, with flexible thinking control that lets you tune reasoning depth per request. It processes text and image input, with video input available in Beta, making it well-suited to complex multi-step workloads, autonomous agents, and vision-language applications.

Intelligence	Speed	Input	Output
High	Medium	Text, Image, Video	Text

Central parameters

Description: Compact dense multimodal model from Alibaba Cloud with 27B parameters and an integrated vision encoder. It offers strong reasoning, agentic execution, and image understanding, with video understanding in Beta.

Model identifier: Qwen/Qwen3.8-27B

IONOS CLOUD AI Model Hub Lifecycle and Alternatives

IONOS CLOUD start date	End of Life	Alternative	Successor
27 August 2026	N/A

Origin

Provider	Country	License	Flavor	Release
Alibaba Cloud	China	Apache 2.0	Instruct	August 2026

Technology

Context window	Parameters	Quantization	Multilingual	Further details
262k	27B	FP8	Yes	Hugging Face

Modalities

Text	Image	Video	Audio
Input and output	Input	Input (beta)	Not supported

Important: Video input is in Beta. Video understanding is available for evaluation and may change without notice. Behaviour, latency, and token accounting for video are not yet covered by production service levels. Text and image input are generally available.

Note: Video content is provided as a public URL using the video_url content type so the AI Model Hub can fetch it. Additionally, you can include up to four images in a single request.

Endpoints

Chat completions	Embeddings	Image generation
v1/chat/completions	Not supported	Not supported

Features

Streaming	Reasoning	Tool calling
Supported	Supported	Supported

Reasoning example

Qwen3.8-27B operates in thinking mode by default, generating an internal reasoning trace before producing its final response. It makes the model well-suited for complex, multi-step reasoning and agentic tasks.

Unlike Qwen3.5-9B, this model supports configurable reasoning depth through reasoning_effort. It accepts the following four values, and the default is medium:

Value	Behaviour
xhigh	For complex tasks demanding thorough analysis
medium (default)	Balances accuracy and speed
low	Efficient reasoning, optimized for speed and cost
none	Deactivates reasoning; returns only the final answer

If you do not send reasoning_effort, the request uses medium. Set the value explicitly whenever you need a different trade-off between answer quality, latency, and token cost.

Note: The reasoning trace counts toward max_tokens. If max_tokens is set too low, reasoning can consume the entire token budget, leaving no tokens for the final answer. Activating reasoning also increases response latency, since the model generates the reasoning trace in addition to the final answer.

Note: In multi-turn agentic tasks, a lower reasoning effort does not always reduce overall completion time. Each turn may respond faster, but insufficient analysis can cause more failures and retries. That can increase both total latency and token consumption.

Request

{
  "stream": false,
  "model": "Qwen/Qwen3.8-27B",
  "messages": [
    {
      "role": "user",
      "content": "Answer me with one letter. Maybe A."
    }
  ]
}

Response (shortened for readability)

{
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "A",
        "reasoning": "User asks to answer with one letter, maybe A. So respond with a single letter."
      }
    }
  ]
}

Control reasoning depth

To trade some accuracy for lower latency and cost, lower the reasoning effort.

{
  "stream": false,
  "model": "Qwen/Qwen3.8-27B",
  "reasoning_effort": "low",
  "messages": [
    {
      "role": "user",
      "content": "Answer me with one letter. Maybe A."
    }
  ]
}

Deactivate reasoning

To skip the internal reasoning trace and return only the final answer, set "reasoning_effort": "none".

{
  "stream": false,
  "model": "Qwen/Qwen3.8-27B",
  "reasoning_effort": "none",
  "messages": [
    {
      "role": "user",
      "content": "Answer me with one letter. Maybe A."
    }
  ]
}

Usage example

Chat completions

The following example demonstrates how to use Qwen3.8-27B for everyday conversational tasks, such as drafting a concise product description.

API Endpoint: POST https://openai.inference.de-txl.ionos.com/v1/chat/completions

Request:

{
  "model": "Qwen/Qwen3.8-27B",
  "messages": [
    {
      "role": "system",
      "content": "You are a helpful marketing assistant. Write clear, concise product copy."
    },
    {
      "role": "user",
      "content": "Write a one-sentence product description for a stainless steel water bottle that keeps drinks cold for 24 hours."
    }
  ],
  "temperature": 0.5,
  "max_tokens": 1000
}

Response:

{
  "id": "chatcmpl-890",
  "object": "chat.completion",
  "created": 1677652289,
  "model": "Qwen/Qwen3.8-27B",
  "choices": [{
    "index": 0,
    "message": {
      "role": "assistant",
      "content": "Stay refreshed all day with our double-walled stainless steel water bottle, engineered to keep your drinks ice-cold for a full 24 hours.",
      "reasoning": "The user wants a single-sentence product description for an insulated stainless steel water bottle highlighting 24-hour cold retention. I will keep it to one sentence and emphasize the key benefit."
    },
    "finish_reason": "stop"
  }],
  "usage": {
    "prompt_tokens": 41,
    "completion_tokens": 33,
    "total_tokens": 74
  }
}

Vision model example

{
  "model": "Qwen/Qwen3.8-27B",
   "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "text",
            "text": "What do you see in this image?"
          },
          {
            "type": "image_url",
            "image_url": {
              "url": "https://www.ionos.com/favicon.ico?v2",
              "detail": "auto"
            }
          }
        ]
      }
    ],
    "temperature": 0.7,
    "max_tokens": 1000
  }

Video input example

Qwen3.8-27B also accepts video input. Pass a clip with the video_url content type; the AI Model Hub fetches the file, so the URL must be public.

Video input is currently in Beta; see the note in Modalities.

{
  "model": "Qwen/Qwen3.8-27B",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "text",
          "text": "Describe what happens in this video."
        },
        {
          "type": "video_url",
          "video_url": {
            "url": "https://example.com/sample-clip.mp4"
          }
        }
      ]
    }
  ],
  "temperature": 0.7,
  "max_tokens": 1000
}

Stream chat completions

To receive responses in real time, you can use streaming. When streaming is selected, usage statistics are not included by default. To obtain usage information in the final stream chunk, you must explicitly set "stream_options": {"include_usage": true}.

API Endpoint: POST https://openai.inference.de-txl.ionos.com/v1/chat/completions

Request:

{
  "model": "Qwen/Qwen3.8-27B",
  "messages": [
    {
      "role": "user",
      "content": "Compose a poem about the sea."
    }
  ],
  "stream": true,
  "stream_options": {
    "include_usage": true
  }
}

Response:

...
data: {"choices":[],"created":1769419547,"id":"chatcmpl-3d0323bb-67f8-41c5-b757-8fc5a48230f4","model":"Qwen/Qwen3.8-27B","object":"chat.completion.chunk","usage":{"completion_tokens":26,"prompt_tokens":78,"total_tokens":104}}

data: [DONE]

Rate limits

Rate limits ensure fair usage and reliable access to the AI Model Hub. In addition to the contract-wide rate limits, no model-specific limits apply.