跳转至

推荐的 BYOK 模型

BYOK 的 AI Model 下拉菜单中列出的所有模型均被视为推荐模型，并可与 BYOK 一起使用。

推荐徽章代表选择建议：

  • High Performance：当您需要最高扫描质量和一致性时的最佳选项。
  • Cost Effective：预算优先时的较低成本选项。
  • 无徽章：标准推荐选项。

按提供商分类的推荐模型

AWS Bedrock

  • anthropic.claude-opus-4-8 (High Performance)
  • anthropic.claude-opus-4-7
  • anthropic.claude-sonnet-4-6 (High Performance)
  • anthropic.claude-opus-4-6-v1

Azure AI Foundry

  • gpt-5.5 (High Performance)
  • moonshot.kimi-k2-6 (Cost Effective)
  • xai.grok-4-3 (High Performance)
  • deepseek.v4-pro (High Performance)
  • anthropic.claude-v4-7 (High Performance)
  • anthropic.claude-v4-6 (High Performance)

Anthropic

  • claude-opus-4-8 (High Performance)
  • claude-opus-4-7
  • claude-opus-4-6 (High Performance)
  • claude-sonnet-4-6 (High Performance)
  • claude-opus-4-5-20251101
  • claude-sonnet-4-5-20250929

OpenAI

  • gpt-5.5 (High Performance)

Google

  • gemini-3.5-flash (High Performance)
  • gemini-3.1-pro-preview (High Performance)
  • gemini-3.1-flash-lite-preview (Cost Effective)

Z.AI

  • glm-5.2
  • glm-5.1

xAI

  • grok-4.5 (High Performance)
  • grok-4.3

Moonshot AI

  • kimi-k2.7-code (High Performance)
  • kimi-k2.7-code-highspeed
  • kimi-k2.6
  • kimi-k2.5

Deepseek

  • deepseek-v4-pro
  • deepseek-v4-flash (Cost Effective)
  • deepseek-3.2

OpenRouter

  • deepseek/deepseek-v4-pro
  • deepseek/deepseek-v4-flash (Cost Effective)
  • deepseek/deepseek-v3.2
  • xiaomi/mimo-v2.5-pro
  • anthropic/claude-fable-5 (High Performance)
  • anthropic/claude-opus-4.8 (High Performance)
  • anthropic/claude-opus-4.6 (High Performance)
  • anthropic/claude-sonnet-4.6 (High Performance)
  • anthropic/claude-opus-4.5
  • anthropic/claude-sonnet-4.5
  • openai/gpt-5.5 (High Performance)
  • google/gemini-3.5-flash (High Performance)
  • google/gemini-3.1-pro-preview (High Performance)
  • google/gemini-3.1-flash-lite-preview (Cost Effective)
  • z-ai/glm-5.2
  • z-ai/glm-5.1
  • x-ai/grok-4.20
  • minimax/minimax-m3 (Cost Effective)
  • minimax/minimax-m2.7 (Cost Effective)
  • minimax/minimax-m2.5 (Cost Effective)
  • moonshotai/kimi-k2.7-code (High Performance)
  • moonshotai/kimi-k2.6 (Cost Effective)
  • moonshotai/kimi-k2.5
  • qwen/qwen3.7-max (High Performance)
  • qwen/qwen3.7-plus (Cost Effective)

注意事项

  • 为获得更好的结果，请尽可能在 Agentic Deep Scans 中优先选择标记为 High Performance 的模型。
  • 模型的可用性取决于您的提供商账户和订阅计划。