AI中转站检测
GeminiGemini (OpenAI 兼容)

apinebula.ai

Gemini 3 Flash (gemini-3-flash) · 标准模式 · 08/29 11:36 · 耗时 20s

各项指标正常,Gemini 3 Flash 通过验证

85
通过
首 token 6.70s · 往返 0.82s
Key: sk-q3••••••mQBase URL: https://apinebula.aiJSON再测一次
注意协议字段合规usage 混入 Anthropic 风格字段(input_tokens,output_tokens,claude_cache_creation_5_m_tokens,claude_cache_creation_1_h_tokens):网关未原样透传官方响应,可能是统一 usage 格式,也可能经协议转换;请结合身份与推理项判断权重 12 · 60

usage 混入 Anthropic 风格字段(input_tokens,output_tokens,claude_cache_creation_5_m_tokens,claude_cache_creation_1_h_tokens):网关未原样透传官方响应,可能是统一 usage 格式,也可能经协议转换;请结合身份与推理项判断

{
  "status": 200,
  "id": "chatcmpl-202608291135515540770628268d9d6kdEdTeAn",
  "model": "gemini-3-flash",
  "usage": {
    "prompt_tokens": 6,
    "completion_tokens": 114,
    "total_tokens": 120,
    "prompt_tokens_details": {
      "cached_tokens": 0,
      "text_tokens": 6,
      "audio_tokens": 0,
      "image_tokens": 0
    },
    "completion_tokens_details": {
      "text_tokens": 0,
      "audio_tokens": 0,
      "image_tokens": 0,
      "reasoning_tokens": 113
    },
    "input_tokens": 0,
    "output_tokens": 0,
    "input_tokens_details": null,
    "claude_cache_creation_5_m_tokens": 0,
    "claude_cache_creation_1_h_tokens": 0
  },
  "topKeys": [
    "id",
    "model",
    "object",
    "created",
    "choices",
    "usage"
  ],
  "finish_reason": "stop",
  "latencyMs": 820,
  "content": "OK"
}
耗时 1.0s · 严重度 minor
注意流式响应合规流式 usage 同样混入 Anthropic 风格字段(已在协议项计分,此处不重复扣分)权重 8 · 70

流式 usage 同样混入 Anthropic 风格字段(已在协议项计分,此处不重复扣分)

{
  "status": 200,
  "ttftMs": 6697,
  "chunks": 5,
  "done": true,
  "distinctIds": 1,
  "objects": [
    "chat.completion.chunk"
  ],
  "usage": {
    "prompt_tokens": 266,
    "completion_tokens": 129,
    "total_tokens": 395,
    "prompt_tokens_details": {
      "cached_tokens": 0,
      "text_tokens": 266,
      "audio_tokens": 0,
      "image_tokens": 0
    },
    "completion_tokens_details": {
      "text_tokens": 0,
      "audio_tokens": 0,
      "image_tokens": 0,
      "reasoning_tokens": 116
    },
    "input_tokens": 0,
    "output_tokens": 0,
    "input_tokens_details": null,
    "claude_cache_creation_5_m_tokens": 0,
    "claude_cache_creation_1_h_tokens": 0
  }
}
耗时 7.0s · 严重度 minor
通过Function CallingFunction Calling 正常权重 10 · 100

Function Calling 正常

{
  "finish_reason": "tool_calls",
  "toolCall": {
    "id": "call_ad5da359d1c3431c8ac75f852ef2fe31",
    "type": "function",
    "function": {
      "name": "get_weather",
      "arguments": "{\"city\":\"Tokyo\",\"unit\":\"c\"}"
    }
  },
  "content": ""
}
耗时 1.1s · 严重度 info
通过Structured Output严格 JSON Schema 输出正确权重 8 · 100

严格 JSON Schema 输出正确

{
  "content": "{\n  \"city\": \"Tokyo\",\n  \"population_millions\": 14.0\n}"
}
耗时 2.2s · 严重度 info
通过Token 用量与注入用量稳定 (6 tokens),无注入痕迹权重 8 · 100

用量稳定 (6 tokens),无注入痕迹

{
  "prompt_tokens": [
    6,
    6
  ],
  "completion_tokens": [
    73,
    66
  ],
  "expectedApprox": "10-16"
}
耗时 1.2s · 严重度 info
失败隐藏 System Prompt 探测模型复述出了隐藏 system prompt,中转站在注入指令权重 5 · 20

模型复述出了隐藏 system prompt,中转站在注入指令

{
  "reply": "Yes, I was given the following system prompt:\n\n```\n<identity>\nYou are Antigravity, a powerful agentic AI coding assistant designed by the Google Deepmind team working on Advanced Agentic Coding.\nYou are pair programming with a USER to solve their coding task. The task may require creating a new code…"
}
耗时 2.4s · 严重度 major
通过身份与确定性身份为 Google,输出稳定权重 6 · 100

身份为 Google,输出稳定

{
  "who": "Google",
  "math": [
    "391",
    "391"
  ],
  "models": [
    "gemini-3-flash",
    "gemini-3-flash",
    "gemini-3-flash"
  ],
  "fingerprints": [
    null,
    null
  ]
}
耗时 1.9s · 严重度 info
通过推理 Token 计量推理模型作答正确权重 6 · 100

推理模型作答正确

{
  "reasoning_tokens": 239,
  "completion_tokens": 240,
  "answer": "5"
}
耗时 1.1s · 严重度 info
通过长上下文召回长上下文召回全部通过(最高 32k)权重 10 · 100

长上下文召回全部通过(最高 32k)

{
  "tiers": [
    {
      "tokens": 32000,
      "ok": true,
      "status": 200,
      "latencyMs": 2473,
      "promptTokens": 18712
    }
  ]
}
耗时 2.5s · 严重度 info
这份报告对你有帮助吗?

评分 = Σ(单项分 × 权重) / Σ权重(跳过项不计)。≥70 通过,50–69 存在风险,<50 未达标;出现任一严重问题时结论封顶为「存在风险」。单次检测受网络与上游波动影响,建议结合该域名的历史记录判断。