Skip to content

Throughput

Fastest text models (hosted APIs)

As of October 25, 2026, the fastest hosted text model we track is GPT-5.6 Terra Pro at 21464 output tokens per second, with a normalized speed score of 100.0 out of 100.

Sorted by our normalized output-speed benchmark (higher is faster relative to models we track). Latency and TTFT differ — use the interactive leaderboard’s realtime-chat scenario for a blended view.

  1. 1
    GPT-5.6 Terra Pro
    OpenAI
    100.0
  2. 2
    GPT-6.1 Sol Pro
    OpenAI
    55.3
  3. 3
    GLM 5
    Zhipu AI (GLM)
    53.2
  4. 4
    GPT-6 Astra Pro
    OpenAI
    51.4
  5. 5
    Muse Spark 1.1
    Meta
    6.2
  6. 6
    GPT-5.6 Luna Pro
    OpenAI
    5.7
  7. 7
    GPT-5.6 Sol Pro
    OpenAI
    4.5
  8. 8
    GPT-5.3-Codex
    OpenAI
    4.5
  9. 9
    Muse Spark 1.2
    Meta
    3.0
  10. 10
    GPT-6 Luna Pro
    OpenAI
    2.9
  11. 11
    GPT-6 Sol Pro
    OpenAI
    2.6
  12. 12
    Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)
    Google
    1.9
  13. 13
    Gemini 3.5 Flash
    Google
    1.7
  14. 14
    Muse Spark Open
    Meta
    1.7
  15. 15
    Nemotron 3.5 Super VL
    NVIDIA
    1.6
  16. 16
    MiMo-V2.6-Pro-UltraSpeed
    Xiaomi (MiMo)
    1.5
  17. 17
    Inkling Small
    Thinking Machines
    1.5
  18. 18
    Nemotron 3 Nano 30B A3B
    NVIDIA
    1.4
  19. 19
    Gemini 3.6 Flash
    Google
    1.3
  20. 20
    GPT-4.1 Nano
    OpenAI
    1.2
  21. 21
    Claude Opus 5.5
    Anthropic
    1.1
  22. 22
    GPT-5.4 Mini
    OpenAI
    1.0
  23. 23
    Gemini 3.5 Flash Lite
    Google
    1.0
  24. 24
    Muse Wonder
    Meta
    1.0
  25. 25
    Gemini 2.5 Flash Lite
    Google
    1.0
  26. 26
    Nano Banana 2 (Gemini 3.1 Flash Image)
    Google
    1.0
  27. 27
    o1-mini
    OpenAI
    1.0
  28. 28
    Claude 3.5 Sonnet
    Anthropic
    1.0
  29. 29
    Claude 3.5 Haiku
    Anthropic
    1.0
  30. 30
    Claude 3 Opus
    Anthropic
    1.0
  31. 31
    Claude 3.7 Sonnet
    Anthropic
    1.0
  32. 32
    Gemini 1.5 Pro
    Google
    1.0
  33. 33
    Gemini 1.5 Flash
    Google
    1.0
  34. 34
    Gemini 2.0 Flash
    Google
    1.0
  35. 35
    Grok 2
    xAI
    1.0
  36. 36
    Step 5 Preview
    StepFun
    1.0
  37. 37
    Gemini 3.8 Flash
    Google
    1.0
  38. 38
    Gemini 3.7 Flash
    Google
    0.9
  39. 39
    Qwen3.5-35B-A3B
    Alibaba (Qwen)
    0.9
  40. 40
    GPT-5.4 Nano
    OpenAI
    0.9
  41. 41
    Gemini 3.1 Flash Lite
    Google
    0.9
  42. 42
    Qwen3 235B (Thinking)
    Alibaba (Qwen)
    0.9
  43. 43
    Gemma 3 4B
    Google
    0.8
  44. 44
    Gemini 3.1 Pro
    Google
    0.8
  45. 45
    Nano Banana 2.1
    Google
    0.8
  46. 46
    Qwen3 Next 80B A3B Thinking
    Alibaba (Qwen)
    0.8
  47. 47
    Qwen3.5-Flash
    Alibaba (Qwen)
    0.8
  48. 48
    Voxtral Small 24B 2507
    Mistral
    0.8
  49. 49
    Qwen3.5-122B-A10B
    Alibaba (Qwen)
    0.8
  50. 50
    GPT-5.1-Codex
    OpenAI
    0.8
  51. 51
    Command A+
    Cohere
    0.8
  52. 52
    Muse Spark 1.3
    Meta
    0.8
  53. 53
    GPT-6 Sol
    OpenAI
    0.7
  54. 54
    Codestral 2508
    Mistral
    0.7
  55. 55
    GPT-4.1 Mini
    OpenAI
    0.7
  56. 56
    Gemini 2.5 Flash
    Google
    0.7
  57. 57
    Nova 2 Lite
    Amazon
    0.7
  58. 58
    Nano Banana Pro (Gemini 3 Pro Image)
    Google
    0.7
  59. 59
    GPT-4.1
    OpenAI
    0.7
  60. 60
    Gemini 2.5 Pro
    Google
    0.7
  61. 61
    Seed 1.6 Flash
    ByteDance Seed
    0.7
  62. 62
    Gemini 3 Flash
    Google
    0.7
  63. 63
    Mistral Small 4
    Mistral
    0.7
  64. 64
    o3-mini
    OpenAI
    0.7
  65. 65
    GLM 5.3 FlashX
    Zhipu AI (GLM)
    0.6
  66. 66
    Step 3.7 Flash
    StepFun
    0.6
  67. 67
    GLM 4.5 Air
    Zhipu AI (GLM)
    0.6
  68. 68
    DeepSeek V4.1 Flash
    DeepSeek
    0.6
  69. 69
    Qwen3.6 Flash
    Alibaba (Qwen)
    0.6
  70. 70
    Mistral Small 3.2 24B
    Mistral
    0.6
  71. 71
    Qwen3 Coder 480B A35B
    Alibaba (Qwen)
    0.6
  72. 72
    GPT-5.6 Luna
    OpenAI
    0.6
  73. 73
    GLM-4.7
    Zhipu AI (GLM)
    0.6
  74. 74
    Ministral 3 3B 2512
    Mistral
    0.6
  75. 75
    Nemotron 3.5 Lightning
    NVIDIA
    0.6
  76. 76
    o3
    OpenAI
    0.6
  77. 77
    Mistral Medium 3.5
    Mistral
    0.6
  78. 78
    Qwen3.6 35B A3B
    Alibaba (Qwen)
    0.6
  79. 79
    DeepSeek V4 Flash 0731
    DeepSeek
    0.6
  80. 80
    Mixtral 8x22B
    Mistral
    0.6
  81. 81
    GLM 4.5V
    Zhipu AI (GLM)
    0.6
  82. 82
    Mistral Medium 3.1
    Mistral
    0.6
  83. 83
    Inkling
    Thinking Machines
    0.6
  84. 84
    GLM 5 Turbo
    Zhipu AI (GLM)
    0.5
  85. 85
    Qwen3.5-9B
    Alibaba (Qwen)
    0.5
  86. 86
    Gemini 2.5 Pro (Max Thinking)
    Google
    0.5
  87. 87
    Grok Build 0.1
    xAI
    0.5
  88. 88
    Mistral Medium 3
    Mistral
    0.5
  89. 89
    Qwen3 Coder Flash
    Alibaba (Qwen)
    0.5
  90. 90
    Devstral 2 2512
    Mistral
    0.5
  91. 91
    GPT-5.2-Codex
    OpenAI
    0.5
  92. 92
    Saba
    Mistral
    0.5
  93. 93
    GPT-4o mini
    OpenAI
    0.5
  94. 94
    Nemotron 3 Ultra
    NVIDIA
    0.5
  95. 95
    GLM 4.5
    Zhipu AI (GLM)
    0.5
  96. 96
    GPT-5.1
    OpenAI
    0.5
  97. 97
    Qwen3.5-27B
    Alibaba (Qwen)
    0.5
  98. 98
    GPT-4o
    OpenAI
    0.5
  99. 99
    GLM-4.6
    Zhipu AI (GLM)
    0.5
  100. 100
    DeepSeek V4 Flash Vision Exp
    DeepSeek
    0.5
  101. 101
    GPT-5.5
    OpenAI
    0.5
  102. 102
    GLM 5.1
    Zhipu AI (GLM)
    0.5
  103. 103
    Claude Sonnet 5.5
    Anthropic
    0.4
  104. 104
    GLM 5.2
    Zhipu AI (GLM)
    0.4
  105. 105
    Ministral 3 14B 2512
    Mistral
    0.4
  106. 106
    Qwen3 Coder Next
    Alibaba (Qwen)
    0.4
  107. 107
    o1
    OpenAI
    0.4
  108. 108
    Ministral 3 8B 2512
    Mistral
    0.4
  109. 109
    Seed-2.0-Mini
    ByteDance Seed
    0.4
  110. 110
    Llama 3.1 70B Instruct
    Meta
    0.4
  111. 111
    GPT-5.6 Sol
    OpenAI
    0.4
  112. 112
    MiniMax M3.1 Flash
    MiniMax
    0.4
  113. 113
    Claude Opus 5
    Anthropic
    0.4
  114. 114
    DeepSeek V4 Pro 0423
    DeepSeek
    0.4
  115. 115
    GPT-5.6 Terra
    OpenAI
    0.4
  116. 116
    Hy3
    Tencent (Hunyuan)
    0.4
  117. 117
    Muse Glimmer 30B
    Meta
    0.4
  118. 118
    Qwen3.8 Max Prime
    Alibaba (Qwen)
    0.4
  119. 119
    MiniMax M2-her
    MiniMax
    0.4
  120. 120
    GLM 5.3
    Zhipu AI (GLM)
    0.4
  121. 121
    GLM 5V Turbo
    Zhipu AI (GLM)
    0.4
  122. 122
    Step 3.5 Flash
    StepFun
    0.3
  123. 123
    Grok 4.7
    xAI
    0.3
  124. 124
    Grok 4.20
    xAI
    0.3
  125. 125
    gpt-oss-20b
    OpenAI
    0.3
  126. 126
    Qwen3.7 Flash
    Alibaba (Qwen)
    0.3
  127. 127
    Qwen3.8 Omni Flash
    Alibaba (Qwen)
    0.3
  128. 128
    Qwen3 14B
    Alibaba (Qwen)
    0.3
  129. 129
    Claude Haiku 4.5
    Anthropic
    0.3
  130. 130
    GPT-5.4
    OpenAI
    0.3
  131. 131
    Qwen3.7 Max
    Alibaba (Qwen)
    0.3
  132. 132
    GPT-5.2
    OpenAI
    0.3
  133. 133
    MiniMax M3
    MiniMax
    0.3
  134. 134
    Hy4 preview
    Tencent (Hunyuan)
    0.3
  135. 135
    Seed-2.0-Code
    ByteDance Seed
    0.3
  136. 136
    Qwen3 VL 30B A3B Thinking
    Alibaba (Qwen)
    0.3
  137. 137
    GLM 5.3 Prime
    Zhipu AI (GLM)
    0.3
  138. 138
    Claude Sonnet 5
    Anthropic
    0.3
  139. 139
    Qwen3.6 27B
    Alibaba (Qwen)
    0.3
  140. 140
    GLM 4.6V
    Zhipu AI (GLM)
    0.3
  141. 141
    Mistral Large 4
    Mistral
    0.3
  142. 142
    GPT-6 Luna
    OpenAI
    0.3
  143. 143
    Grok 4.6
    xAI
    0.3
  144. 144
    Qwen3.8 27B
    Alibaba (Qwen)
    0.3
  145. 145
    Gemini 4 Argon
    Google
    0.3
  146. 146
    GLM 5.3 Flash
    Zhipu AI (GLM)
    0.3
  147. 147
    Qwen3 Next 80B A3B Instruct
    Alibaba (Qwen)
    0.3
  148. 148
    Qwen3.7 Plus
    Alibaba (Qwen)
    0.3
  149. 149
    Qwen3.5 Plus 2026-02-15
    Alibaba (Qwen)
    0.3
  150. 150
    Gemma 3 12B
    Google
    0.2
  151. 151
    Qwen3.5 397B A17B
    Alibaba (Qwen)
    0.2
  152. 152
    Grok 4.3
    xAI
    0.2
  153. 153
    Kimi K2.7 Code
    Moonshot (Kimi)
    0.2
  154. 154
    Qwen3.6 Plus
    Alibaba (Qwen)
    0.2
  155. 155
    Qwen3.5 Plus 2026-04-20
    Alibaba (Qwen)
    0.2
  156. 156
    Mistral Large 3 2512
    Mistral
    0.2
  157. 157
    Mistral Small 3.1 24B
    Mistral
    0.2
  158. 158
    Claude Opus 4.8
    Anthropic
    0.2
  159. 159
    Grok 4.5
    xAI
    0.2
  160. 160
    Command A
    Cohere
    0.2
  161. 161
    Qwen-Plus
    Alibaba (Qwen)
    0.2
  162. 162
    Grok 3
    xAI
    0.2
  163. 163
    MiniMax M2.5
    MiniMax
    0.2
  164. 164
    MiniMax M2.1
    MiniMax
    0.2
  165. 165
    GPT-5
    OpenAI
    0.2
  166. 166
    Mistral Large 2
    Mistral
    0.2
  167. 167
    Mistral Small 3
    Mistral
    0.2
  168. 168
    Nano Banana (Gemini 2.5 Flash Image)
    Google
    0.2
  169. 169
    GPT-5 mini
    OpenAI
    0.2
  170. 170
    GPT-6.1 Sol
    OpenAI
    0.2
  171. 171
    Qwen3.8 2.4T A95B
    Alibaba (Qwen)
    0.2
  172. 172
    MiniMax M2
    MiniMax
    0.2
  173. 173
    Qwen2.5 VL 72B Instruct
    Alibaba (Qwen)
    0.2
  174. 174
    Claude Sonnet 4 (Thinking)
    Anthropic
    0.2
  175. 175
    Claude Opus 4.7
    Anthropic
    0.2
  176. 176
    o4-mini
    OpenAI
    0.2
  177. 177
    Gemma 4 31B
    Google
    0.2
  178. 178
    GPT-6 Astra
    OpenAI
    0.2
  179. 179
    GPT-5.4 Image 2
    OpenAI
    0.2
  180. 180
    GLM 4.7 Flash
    Zhipu AI (GLM)
    0.2
  181. 181
    Seed 1.6
    ByteDance Seed
    0.2
  182. 182
    Kimi K2
    Moonshot (Kimi)
    0.2
  183. 183
    GPT-5 nano
    OpenAI
    0.2
  184. 184
    Kimi K2 0905
    Moonshot (Kimi)
    0.2
  185. 185
    MiniMax M1
    MiniMax
    0.2
  186. 186
    Qwen3 30B A3B
    Alibaba (Qwen)
    0.2
  187. 187
    Qwen3.8 Flash
    Alibaba (Qwen)
    0.2
  188. 188
    Nemotron 3 Super
    NVIDIA
    0.2
  189. 189
    Claude Sonnet 4
    Anthropic
    0.2
  190. 190
    Grok 4
    xAI
    0.2
  191. 191
    GPT-5.1-Codex-Mini
    OpenAI
    0.2
  192. 192
    Seed 2.1 Turbo
    ByteDance Seed
    0.2
  193. 193
    Hunyuan A13B Instruct
    Tencent (Hunyuan)
    0.2
  194. 194
    GPT-4 Turbo
    OpenAI
    0.1
  195. 195
    Seed-2.0-Lite
    ByteDance Seed
    0.1
  196. 196
    Qwen2.5 72B Instruct
    Alibaba (Qwen)
    0.1
  197. 197
    Claude Opus 4.5
    Anthropic
    0.1
  198. 198
    Qwen3.8 Max (0902)
    Alibaba (Qwen)
    0.1
  199. 199
    Kimi K3
    Moonshot (Kimi)
    0.1
  200. 200
    GPT-5.1-Codex-Max
    OpenAI
    0.1
  201. 201
    MiniMax M2.7
    MiniMax
    0.1
  202. 202
    DeepSeek V4 Flash 0423
    DeepSeek
    0.1
  203. 203
    Claude Sonnet 4.6
    Anthropic
    0.1
  204. 204
    gpt-oss-120b
    OpenAI
    0.1
  205. 205
    Claude Opus 4
    Anthropic
    0.1
  206. 206
    Claude Opus 4 (Thinking)
    Anthropic
    0.1
  207. 207
    Kimi K2.6
    Moonshot (Kimi)
    0.1
  208. 208
    MiMo-V2.5
    Xiaomi (MiMo)
    0.1
  209. 209
    Kimi K3.1 Pro
    Moonshot (Kimi)
    0.1
  210. 210
    Claude Opus 4.6
    Anthropic
    0.1
  211. 211
    Qwen3 Coder 30B A3B Instruct
    Alibaba (Qwen)
    0.1
  212. 212
    MiMo-V2.6-Flash
    Xiaomi (MiMo)
    0.1
  213. 213
    Claude Sonnet 4.5
    Anthropic
    0.1
  214. 214
    Qwen3 VL 30B A3B Instruct
    Alibaba (Qwen)
    0.1
  215. 215
    Llama 3.1 405B Instruct
    Meta
    0.1
  216. 216
    Gemma 4 26B A4B
    Google
    0.1
  217. 217
    MiMo-V2.5-Pro
    Xiaomi (MiMo)
    0.1
  218. 218
    MiMo-V2.6-Pro
    Xiaomi (MiMo)
    0.1
  219. 219
    Gemma 3 27B
    Google
    0.1
  220. 220
    Qwen3 30B A3B Instruct 2507
    Alibaba (Qwen)
    0.1
  221. 221
    DeepSeek V3
    DeepSeek
    0.1
  222. 222
    Llama 4 Scout
    Meta
    0.1
  223. 223
    Qwen3 32B
    Alibaba (Qwen)
    0.1
  224. 224
    DeepSeek V3 (Thinking)
    DeepSeek
    0.1
  225. 225
    MiniMax-01
    MiniMax
    0.1
  226. 226
    DeepSeek V3.1 Terminus
    DeepSeek
    0.1
  227. 227
    Qwen3 235B
    Alibaba (Qwen)
    0.1
  228. 228
    DeepSeek R1
    DeepSeek
    0.1
  229. 229
    Qwen3 VL 235B A22B Instruct
    Alibaba (Qwen)
    0.0
  230. 230
    Llama 3.3 70B Instruct
    Meta
    0.0
  231. 231
    DeepSeek V3.1
    DeepSeek
    0.0
  232. 232
    Qwen3 VL 8B Instruct
    Alibaba (Qwen)
    0.0
  233. 233
    Llama 4 Maverick
    Meta
    0.0
Realtime scenario Leaderboard

Speed figures are aggregated from observable public sources; see AI Model Analyzer About for methodology.