Skip to main content

Qwen3.6 27B

Qwen's latest-generation 27B — strong general reasoning with built-in thinking style.

Overview​

  • model_name: hf-Qwen-Qwen3.6-27B (HuggingFace import, currently free)
  • Source: Qwen/Qwen3.6-27B
  • Scale: 27B parameters (Qwen3.5/3.6 architecture generation)
  • Served on: GPU via vLLM, tensor-parallel across 2× Intel Arc Pro B70 (32GB)
  • Status: ready (verified end-to-end through the public inference route, 2026-08-03)

When to use​

✅ High-quality general assistant work above the 7-14B class: analysis, long-form writing, multilingual tasks, and step-by-step reasoning (the model thinks before answering by default).

❌ Latency-sensitive chat — a 27B dense model on TP2 trades speed for quality; use Qwen 2.5 7B for fast turns.

Live specs​

curl "https://api.colabhive.com/api/builder/v1/endpoints?visibility=public&search=Qwen3.6"

Quick start​

curl -X POST "https://api.colabhive.com/api/builder/v1/endpoints/<endpoint_id>/infer" \
  -H "X-API-Key: $COLABHIVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"input": {"messages": [{"role": "user", "content": "Explain the CAP theorem with a practical example."}], "max_tokens": 600}}'