高效轻量化MoE模型，总参284B，激活13B，原生支持百万超长上下文能力。推理速度快、延迟低、调用成本低廉，综合能力均衡，主打高并发、轻量化任务，适合日常对话、内容创作、基础 RAG、批量文案处理等普惠刚需场景。
curl https://api.yourplatform.com/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "undefined",
"messages": [
{"role": "user", "content": "Hello!"}
],
"stream": true
}'