Qwen3.8-27B is Alibaba's open-weight 27B dense model, released under Apache-2.0 and self-hosted on OrcaRouter's own infrastructure. This deployment is text-in, text-out, with a 64K-token context window (65,536 positions) built on the Qwen3.5 architecture - 64 layers, 5120 hidden size. Despite its 27B size it posts unusually strong agentic and coding results on Qwen's own model card, including 90.3 on LiveCodeBench v6, 89.2 on GPQA Diamond, and 79.0 on QwenSWEBench - competitive with far larger closed models on several axes. It supports native tool calling, structured outputs, reasoning mode, and the full sampling surface. Because the weights are open and we run them ourselves, there is no per-token vendor cost to pass through. That makes Qwen3.8-27B a strong default for high-volume text workloads: coding agents, long-document analysis, structured extraction, and agentic automation.
OpenAI-compatible — keep the SDK you already use
https://api.orcarouter.ai/v1import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.orcarouter.ai/v1",
api_key=os.environ["ORCAROUTER_API_KEY"],
)
response = client.chat.completions.create(
model="qwen/qwen3.8-27b-free",
messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)frequency_penaltyinclude_reasoninglogprobsmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_p| Per request | $0 |
| Billing | Model usage is never charged to your balance |
| Over the limit | HTTP 429 when a limit is hit |
Need it without the limits? qwen/qwen3.8-27b·How free-tier limits work →
What developers are saying this week
GET /api/public/models/qwen/qwen3.8-27b-freeOpen @misc{orcarouter_qwen3_8_27b_free,
title = {Qwen3.8 27B (free) API},
author = {Qwen},
year = {2026},
howpublished = {OrcaRouter},
url = {https://www.orcarouter.ai/models/qwen/qwen3.8-27b-free}
}Qwen. (2026). Qwen3.8 27B (free) API. OrcaRouter. https://www.orcarouter.ai/models/qwen/qwen3.8-27b-free