Qwen3.8 27B is available on EmpirioLabs. It is Alibaba's open 27B multimodal reasoner from the Qwen3.8 family, with a 262,144-token context, text, image, and video input, function calling, strict structured output, and thinking on by default. Point an OpenAI-compatible client at EmpirioLabs and set model to qwen3-8-27b.
What Qwen3.8 27B is good at
This is the dense 27B member of Qwen3.8, not the hosted Max tier. It fits work that needs a long context and multimodal input without the Cloud extras: coding and analysis with screenshots or clips attached, tool-using agents, and requests that should return exactly the JSON Schema you send. Thinking is on unless you turn it off, so short latency-sensitive calls should set enable_thinking to false.
How to call Qwen3.8 27B
Use the OpenAI-compatible chat completions endpoint:
curl https://api.empiriolabs.ai/v1/chat/completions \
-H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \
-H 'Content-Type: application/json' \
-d '{
"model": "qwen3-8-27b",
"enable_thinking": true,
"reasoning_effort": "xhigh",
"messages": [
{
"role": "user",
"content": "Summarize this screenshot and list the three actions to take next."
}
]
}'The same model id works on /v1/responses, the Anthropic-shaped /v1/messages, legacy /v1/completions, and the Google-compatible /v1beta/models/qwen3-8-27b:generateContent route. Alternate ids qwen3.8-27b, qwen/qwen3-8-27b, and qwen/qwen3.8-27b resolve to the same model.
Thinking, vision, and structured output
Reasoning is controlled by enable_thinking and reasoning_effort (xhigh, medium, low). There is no token thinking budget on this slug. Attach images as standard image_url parts, including more than one image in a turn, and send video the same way the playground does. For machine-readable results, use JSON mode or a strict JSON Schema. Optional web search powered by Linkup can retrieve recent sources for your latest message when you enable it.
This slug serves the open 256K context. Hosted Cloud 1M context and official built-in provider tools are not part of this model.
Pricing
Qwen3.8 27B uses pay-as-you-go token pricing with a steeply discounted rate for cached input and no per-seat fees. Current rates are on the model page and pricing.
Try it
Open the playground, read the API docs, or browse all models and pricing.
