Loading
RedRouterRedRouter

Qwen 3.8 27B RR

qwen3.8-27b-rrReduced refusalsAvailable

Alibaba's mid-size Qwen 3.8 27B, modified to reduce refusals. Balances quality and cost for chat, writing, and multilingual work.

Specs and pricing

Model IDqwen3.8-27b-rr
Context1,010,000 tokens
Max output8,192 tokens
Default output1,024 tokens
Input price$0.55 per 1M tokens
Cached input price$0.08 per 1M tokens
Output price$2.40 per 1M tokens
API compatibilityOpenAI Chat Completions, OpenAI Responses, Anthropic Messages
Tool callingYes
JSON response formatYes

Quickstart

Set REDROUTER_API_KEY to a key from API Keys.

Base URL
https://redrouter.org/v1
Model ID
qwen3.8-27b-rr
curl 'https://redrouter.org/v1/chat/completions' \
  -H "Authorization: Bearer $REDROUTER_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{"messages":[{"content":"Hello","role":"user"}],"model":"qwen3.8-27b-rr"}'

Supported parameters

Other parameters are rejected with HTTP 400. The user field is accepted for compatibility and is not forwarded.

modelmessagesstreamstream_optionsmax_tokensmax_completion_tokenstemperaturetop_pstoppresence_penaltyfrequency_penaltyseedtoolstool_choiceparallel_tool_callsresponse_formatuserreasoning_effortreasoningenable_thinkingthinkingthinking_budgetchat_template_kwargsoutput_config

Request fields and limits

Setup notes and limits: Qwen 3.8 27B RR on the RedRouter blog.

Deployment

Routing
Your request runs on the model you named. RedRouter does not substitute a different model.

Model version

As reported by the hosting provider. RedRouter has not independently verified the served weights.

Weights
Not disclosed by host
Revision
Not disclosed by host
License
Not disclosed by host
Quantization
FP8
Tokenizer
Not disclosed by host
Chat template
Not disclosed by host