Qwen3.8-27B

POST /v1/chat/completions model: Qwen3.8-27B

Specification

Architecture	Qwen3_5ForConditionalGeneration, model_type = qwen3_5
Layers	64
Hidden dimension	5,120
Attention	24 query heads, 4 KV heads (GQA), head dim 256
Intermediate dimension	17,408
Vocabulary	248,320
Vision tower	27 layers, width 1,152, 16 heads, patch 16, spatial merge 2
Precision	bfloat16 — unquantised

At a glance

Context length	131,072 tokens
Input modalities	text + images
Streaming	✅
Tool calling	✅
JSON output	✅ json_object
Thinking	✅ — on by default
Inference engine	vLLM
Stability	experimental

Thinking

Thinking is on unless you turn it down; omitting reasoning_effort falls through to xhigh.

Value
omitted	falls through to xhigh
low	✅
medium	✅
xhigh	✅ longest tier
high minimal max	❌ 400

This model’s top tier is xhigh, not high

reasoning_effort: "high" is rejected:

Unexpected reasoning effort high. Supported types are xhigh (default), medium, and low.

Send xhigh for the longest thinking, or omit the parameter.

Thinking text arrives in choices[0].message.reasoning. This model does not return usage.completion_tokens_details, so thinking tokens are not reported separately; they are included in usage.completion_tokens.

Thinking draws on the same max_tokens budget as the answer. If content comes back empty, raise max_tokens or send reasoning_effort: "low".

messages

A system message is optional, but there can be at most one and it must come first. A second one, or one after a user turn, is rejected:

System message must be at the beginning.

Both system and developer are accepted as the role name.

Image input

Send an image_url content part. Both data: URLs and https:// URLs are accepted.

{
  "model": "Qwen3.8-27B",
  "messages": [{
    "role": "user",
    "content": [
      { "type": "image_url", "image_url": { "url": "data:image/png;base64,iVBORw0KGgo..." } },
      { "type": "text", "text": "What does this image say?" }
    ]
  }]
}

Image usage is reported at usage.prompt_tokens_details.multimodal_tokens.image.

Limits

Context window	131,072 tokens, prompt plus output
JSON output	response_format: {"type": "json_object"}
Thinking tiers	low, medium, xhigh

max_tokens is checked against the whole window:

max_tokens=999999 cannot be greater than max_model_len=max_total_tokens=131072.

Example

curl https://developer.amd.com.cn/radeon/api/v1/chat/completions \
  -H "Authorization: Bearer $RADEON_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "Qwen3.8-27B",
    "reasoning_effort": "low",
    "max_tokens": 512,
    "messages": [
      { "role": "system", "content": "Answer in one sentence." },
      { "role": "user", "content": "Why is the sky blue?" }
    ]
  }'