Maas-Model-platform/backend/parsed_json/qwen3.5-plus.json

{
  "model_code": "qwen3.5-plus",
  "model_intro": "Qwen3.5原生视觉理解系列Plus模型，基于混合架构，融合了链式注意力机制与视觉融合专业模型，实现了更高的推理效率。在多任务评测中，3.5系列展现出与当前顶尖前沿模型相当甚至更优的性能表现，模型效果在通用文本与多模态方面相比3系列实现显著提升。",
  "model_tags": ["Qwen3.5", "Plus", "语言模型", "通义千问", "视觉理解"],
  "model_capabilities": {
    "input_modalities": ["text", "image"],
    "output_modalities": ["text"],
    "features": {
      "text": true,
      "image": true,
      "internet_search": true,
      "function_calling": true,
      "structured_output": true
    }
  },
  "model_pricing": {
    "input": 0.8,
    "output": 2.0,
    "image_input": 1.8,
    "unit": "元/千tokens"
  },
  "model_limits": {
    "max_input_length": "128K tokens",
    "max_output_length": "4K tokens",
    "rpm": "100",
    "tpm": "100000"
  }
}