专项榜

本站 · 编码 Top10

本站编码分 Top10, 与一览表编码 tab 同源 (同字段/同去重键, 组内取该项最高分)

模型一览表 →
  1. #1
    Claude Mythos 5.1
    95.3
  2. #2
    Claude Fable 5.1
    94.8
  3. #3
    Claude Opus 5
    86.3
  4. #4
    Ornith-1.5-397B
    85.2
  5. #5
    OpenAI: GPT-5.3-Codex
    84.2
  6. #6
    Ornith-1.0-397B
    81.4
  7. #7
    Anthropic: Claude Fable 5.1
    80.5
  8. #8
    Anthropic: Claude Opus 4.8
    80.4
  9. #9
    Anthropic: Claude Opus 4.5
    79.8
  10. #10
    Anthropic: Claude Opus 4.6
    79.7

本站 · 推理 Top10

本站推理分 Top10, 与一览表推理 tab 同源 (同字段/同去重键, 组内取该项最高分)

模型一览表 →
  1. #1
    GPT-6 Astra
    90.6
  2. #2
    Sakana Fugu-Ultra v1.1
    89.9
  3. #3
    GPT-5.6 Cyber
    88.7
  4. #4
    Google: Gemini 3.1 Pro Preview
    88.5
  5. #5
    OpenAI: GPT-6 Astra
    88.4
  6. #6
    Google: Gemini 3.7 Flash
    88.3
  7. #7
    OpenAI: GPT-5.6 Sol Pro
    87.6
  8. #8
    SpaceXAI: Grok 4.6
    86.9
  9. #9
    Google: Gemini 3.6 Flash
    86.3
  10. #10
    OpenAI: GPT-5.6 Sol
    85.1

OpenRouter · 周调用榜

OpenRouter 周 token 调用量 Top10（流行度，非能力）

查看原榜 →
  1. #1
    OpenAI: GPT-5.6 Luna
    100
  2. #2
    Tencent: Hy4 preview
    88.9
  3. #3
    DeepSeek: DeepSeek V4 Flash 0731
    77.8
  4. #4
    Z.ai: GLM 5.3 Flash
    66.7
  5. #5
    Xiaomi: MiMo-V2.5
    55.6
  6. #6
    DeepSeek: DeepSeek V4.1 Flash
    44.4
  7. #7
    DeepSeek: DeepSeek V4 Flash 0423
    33.3
  8. #8
    Tencent: Hy3
    22.2
  9. #9
    NVIDIA: Nemotron 3 Ultra
    11.1
  10. #10
    Google: Gemini 3.8 Flash
    0

AA · 最强智能

Artificial Analysis 智能指数 Top10（能力，越高越强）

查看原榜 →
  1. #1
    Anthropic: Claude Fable 5.1
    53.4
  2. #2
    Claude Fable 5.1 (Adaptive Reasoning, Xhigh Effort, Default Fallback)
    53.2
  3. #3
    OpenAI: GPT-6 Astra
    52.8
  4. #4
    GPT-6 Astra (xhigh)
    52.5
  5. #5
    Claude Fable 5.1 (Adaptive Reasoning, High Effort, Default Fallback)
    51.2
  6. #6
    GPT-6 Astra (high)
    51
  7. #7
    Claude Opus 5
    50.7
  8. #8
    Claude Opus 5 (Adaptive Reasoning, Xhigh Effort)
    49.7
  9. #9
    Anthropic: Claude Fable 5
    49.7
  10. #10
    GPT-6 Astra (medium)
    49.7

AA · 最强编码

Artificial Analysis 编码指数 Top10

查看原榜 →
  1. #1
    Anthropic: Claude Fable 5.1
    81.6
  2. #2
    Claude Fable 5.1 (Adaptive Reasoning, Xhigh Effort, Default Fallback)
    80.7
  3. #3
    Claude Fable 5.1 (Adaptive Reasoning, High Effort, Default Fallback)
    79.1
  4. #4
    gpt-5.6-sol-xhigh
    78.3
  5. #5
    Claude Opus 5
    78
  6. #6
    OpenAI: GPT-5.6 Sol
    77.4
  7. #7
    GPT-5.6 Sol (high)
    77.2
  8. #8
    Claude Fable 5.1 (Adaptive Reasoning, Medium Effort, Default Fallback)
    77.1
  9. #9
    GPT-6 Astra (high)
    77.1
  10. #10
    Claude Opus 5 (Adaptive Reasoning, Xhigh Effort)
    77

AA · 最快模型

Artificial Analysis 中位数输出速度 Top10（tok/s，越高越快）

查看原榜 →
  1. #1
    Celeris-1
    1412 tok/s
  2. #2
    Inception: Mercury 2
    807 tok/s
  3. #3
    HyperNova 60B 2605 (high, based on gpt-oss-120b)
    361 tok/s
  4. #4
    Gemini 2.5 Flash-Lite (Reasoning)
    334 tok/s
  5. #5
    Google: Gemini 3.5 Flash Lite
    320 tok/s
  6. #6
    Arcee AI: Trinity Large Thinking
    310 tok/s
  7. #7
    inclusionAI: Ling 3.0 Flash
    307 tok/s
  8. #8
    NVIDIA: Nemotron 3.5 Lightning
    288 tok/s
  9. #9
    Google: Gemini 3.7 Flash
    288 tok/s
  10. #10
    Google: Gemini 3.1 Flash Lite Preview
    288 tok/s

AA · 最具性价比

Artificial Analysis 合成性价比 Top10（智能指数 / 平均价格，越高越值）

查看原榜 →
  1. #1
    Agnes 3.0 Flash
    #1
  2. #2
    Llama 3.1 Instruct 8B
    #2
  3. #3
    Agnes 2.5 Pro Beta
    #3
  4. #4
    inclusionAI: Ling 3.0 Flash
    #4
  5. #5
    Tencent: Hy3 preview
    #5
  6. #6
    Gemma 4 E4B (Reasoning)
    #6
  7. #7
    Qwen3.5 4B (Reasoning)
    #7
  8. #8
    DeepSeek V4 Flash (Non-reasoning)
    #8
  9. #9
    HyperNova 60B 2605 (high, based on gpt-oss-120b)
    #9
  10. #10
    Z.ai: GLM 5.3 Flash
    #10

部分榜单数据由 Artificial Analysis 提供 · artificialanalysis.ai