TrafficScent

AI基準評測

統計基準日 · 來源 Artificial Analysis

語言模型 圖像・影片・音訊

Highlights

Artificial Analysis 核心指標

依智慧指數 (Intelligence)、回應速度 (Speed)、每項任務成本 (Cost per Task) 三大核心指標詳細比較

Intelligence (綜合智慧指數)

Updated
1
Opus 5.5
Anthropic
58
2
Fable 5.1
Anthropic
53
3
Astra 6
OpenAI
53
4
Gemini 4 Argon
Google
53
5
Sol 6.1
OpenAI
52
6
Muse Spark 1.3
Meta
48
7
Grok 4.7
SpaceXAI
46
8
MiMo-V2.6-Pro
Xiaomi
46
9
GLM-5.3
Z AI
45
10
Gemini 3.8 Flash
Google
41

Speed (回應生成速度)

1
Gemini 3.8 Flash
Google
237
2
DeepSeek V4.1 Flash
DeepSeek
213
3
Muse Spark 1.3
Meta
143
4
Opus 5.5
Anthropic
92
5
Grok 4.7
SpaceXAI
79
6
GLM-5.3
Z AI
71
7
Fable 5.1
Anthropic
65
8
Astra 6
OpenAI
60
9
Sol 6.1
OpenAI
52
10
MiMo-V2.6-Pro
Xiaomi
41

Cost per Task (任務成本 · CP值)

1
MiMo-V2.6-Pro
Xiaomi
$0.13
2
DeepSeek V4.1 Flash
DeepSeek
$0.27
3
Sol 6.1
OpenAI
$0.72
4
Gemini 3.8 Flash
Google
$1.24
5
Muse Spark 1.3
Meta
$1.60
6
Gemini 4 Argon
Google
$1.99
7
GLM-5.3
Z AI
$2.01
8
Astra 6
OpenAI
$3.26
9
Grok 4.7
SpaceXAI
$3.74
10
Opus 5.5
Anthropic
$5.98

圖像・影片・音訊

Artificial Analysis Arena

依人類盲測偏好投票 (Arena Elo) 與語音辨識錯誤率 (WER) 排名的各領域頂尖模型

圖像 (Arena Elo)

167 個模型中前 10 名

  1. 1
    GPT Image 2.5 Sunburst
    OpenAI
    1197
  2. 2
    GPT Image 2.5 Flare
    OpenAI
    1191
  3. 3
    GPT Image 2
    OpenAI
    1172
  4. 4
    Grok Imagine Image 2.0
    SpaceXAI
    1155
  5. 5
    MAI-Image-2.6
    Microsoft AI
    1151
  6. 6
    Nano Banana 2
    Google
    1125
  7. 7
    Muse Image
    Meta
    1115
  8. 8
    GPT Image 1.5
    OpenAI
    1107
  9. 9
    MAI-Image-2.6-Flash
    Microsoft AI
    1106
  10. 10
    MAI-Image-2.5
    Microsoft AI
    1104

影片 (Arena Elo)

28 個模型中前 10 名

  1. 1
    Wan 3.0
    Alibaba
    1156
  2. 2
    Utopai X
    Utopai Studios
    1148
  3. 3
    Dreamina Seedance 2.5
    ByteDance Seed
    1142
  4. 4
    MiniMax H3
    MiniMax
    1138
  5. 5
    MiniMax H3 Max
    Fal
    1134
  6. 6
    FLUX 3
    Black Forest Labs
    1127
  7. 7
    Dreamina Seedance 2.0
    ByteDance Seed
    1119
  8. 8
    Gemini Omni Flash 1.1
    Google
    1113
  9. 9
    SkyReels V4
    Skywork AI
    1075
  10. 10
    Agnes-Video-2.5
    Sapiens AI
    1058

音訊 (Arena Elo)

94 個模型中前 10 名

  1. 1
    Eleven v4
    ElevenLabs
    1321
  2. 2
    Qwen-Audio-3.1-TTS-Plus
    Alibaba
    1292
  3. 3
    Sonic 3.6
    Cartesia
    1278
  4. 4
    Gemini 3.8 Flash TTS
    Google
    1275
  5. 5
    Qwen-Audio-3.0-TTS-Plus
    Alibaba
    1261
  6. 6
    Realtime TTS-2
    Inworld
    1251
  7. 7
    Simba 3.2
    SpeechifyAI
    1242
  8. 8
    Gemini 3.8 Flash-Lite TTS
    Google
    1242
  9. 9
    Luna TTS
    VUI Labs
    1227
  10. 10
    Breeze TTS 2
    BreezeBlue
    1216