Fastest AI models

192 models · Most output tokens per second. Ranked models only. Method

Ranked models

November 2026

#ModelSpeedIndexOutput / 1M

01Celeris-11,555/s33.7$0.70

02Beam 1.0 501B-A23B345/s60.2—

03LFM2.5-8B-A1B335/s30.6—

04GPT-OSS 20B310/s41.7$0.13

05Mercury 3295/s52.3$0.35

06LFM2.5-24B-A2B280/s39.8$0.70

07Nemotron 3 Nano 30B270/s32.9—

08DeepSeek V4.1 Flash 1017261/s68.5$0.40

09Nemotron 3 Nano Omni 30B A3B228/s38.6—

10LFM3-1.7B218/s41.5$0.20

11Gemini 3 Flash198/s55.0$3.00

12Sarvam 105B191/s36.7—

13LFM2.5-2.6B189/s29.3$0.50

14o3-mini173/s40.0$4.40

15Grok Code Fast 1172/s41.2$1.50

16Command A+161/s40.2$1.50

17Claude Haiku 5.5159/s66.3$0.50

18Muse Ember158/s64.8$0.50

19Muse Spark 1.1152/s64.6$4.25

20Gemini 3.1 Pro151/s62.3$12.00

20Gemini 3.1 Flash Lite Preview151/s49.6$1.50

22Grok Build 0.1147/s52.7$2.00

23Gemini 2.5 Flash-Lite140/s27.1$0.40

24Trinity-Large-Thinking138/s43.5$0.80

24Grok 4.1 Fast138/s33.3$0.50

26Gemini 3.5 Flash137/s64.2$9.00

27Gemini 3.1 Flash-Lite134/s51.9$1.50

28Gemini 3.5 Flash-Lite130/s52.0$2.50

29MiMo-V2-Flash129/s43.4—

29Nova Pro129/s25.6—

31Mercury 2.5125/s46.2$0.15

32Muse Spark 1.2123/s64.9$4.25

33GLM-5.3-Flash122/s63.1$0.078

33o4-mini122/s44.7$4.40

35Gemini 3.6 Flash121/s63.7$3.75

36Solar Pro 3114/s39.2$0.60

36Mistral Small 4114/s37.9$0.60

38Mistral Large 4113/s60.7$2.10

38Qwen3.8-27B113/s59.2$1.88

40Step 3.7 Flash112/s55.3$1.15

41GLM-5.2111/s60.6$6.80

42Gemini 3 Pro109/s57.9$12.00

43Qwen3 235B 2507 (Reasoning)108/s47.4$3.50

44Claude Sonnet 5.5107/s77.2$10.00

45Step 5 Preview105/s68.4$2.70

46Ling 2.6 Flash104/s37.0—

47GLM-5.1102/s56.0$4.40

48Granite 4.2 8B101/s37.6$0.15

49MiniMax M3.1 Flash100/s69.3$0.60

50DeepSeek V3.198/s37.9—

51o197/s44.5$60.00

52GPT-5.4 mini95/s54.5$4.50

52Gemini 2.5 Flash95/s35.0$2.50

54Mistral Medium 394/s27.1$2.00

55Phi-493/s22.4$0.14

56Mistral Medium 3.592/s42.7$7.50

57GPT-5.4 nano91/s50.9$1.25

57Gemini 2.5 Pro91/s46.6$10.00

59Muse Glimmer 30B90/s53.0$1.20

60DeepSeek V4 Pro 042389/s57.6$1.78

61Claude Haiku 4.588/s43.2$5.00

62GLM-5.386/s65.8$1.40

62Muse Spark Open86/s65.1$4.25

64Gemini 3.7 Flash85/s68.5$3.75

64Inkling85/s54.8$4.05

66Ling 3.1 Flash81/s67.2$0.135

66Kimi K2.681/s58.2$3.41

66GPT-5 nano81/s41.9$0.40

69Muse Wonder80/s73.1$20.00

69Inkling-Small80/s55.1$1.20

71MiniMax M379/s55.7$1.20

72Gemini 3 Flash Preview78/s57.5$3.00

72Grok 478/s50.0—

72GPT-5 mini78/s48.4$2.00

75Gemini 3.1 Pro Preview77/s61.5$12.00

76Claude Opus 5.576/s78.6$20.00

76Muse Spark 1.376/s70.1$4.25

76GPT-576/s55.0$10.00

76GPT-5 (high)76/s53.6$10.00

76GPT-5 (medium)76/s53.0$10.00

81Claude Sonnet 575/s64.2$10.00

81Qwen 3.6 Max (preview)75/s57.2$6.16

83DeepSeek V4 Flash Vision Exp74/s60.4$1.32

83GPT-4.174/s38.0$8.00

85Ling 3.0 Flash73/s50.9$0.063

85Claude 3 Haiku73/s22.0$1.25

87Grok 4.2072/s52.8$2.50

87GPT-4.1 mini72/s34.5$1.60

89Qwen3.5-122B-A10B70/s51.9$2.08

89North Mini Code70/s38.0Free

89Mistral Large 270/s27.4$6.00

92Claude Fable 5.169/s76.2$50.00

92Claude Opus 4.869/s67.7$25.00

92Hy3 Preview69/s52.3$0.60

92Qwen3.5-27B69/s51.3$2.60

96Gemini 3.8 Flash68/s67.8$4.13

96Qwen3.7 Max68/s60.5$4.43

98Hy4 preview67/s66.4$2.50

99GPT-5.6 Terra66/s68.9$12.00

99GPT-6 Luna66/s65.5$0.50

99GPT-4o66/s28.7$10.00

102o365/s51.8$8.00

102Qwen3.6-35B-A3B65/s50.6$0.90

102Grok 3 Mini65/s44.3$0.50

105Claude Fable 5.264/s85.6$40.00

105GPT-5.6 Sol64/s73.0$22.00

107Ripple-65B Preview63/s52.3$0.60

107Llama 4 Scout63/s26.8$0.30

109Qwen3 Max62/s45.0$3.90

110GPT-6 Sol61/s72.8$10.00

110DeepSeek V4 Flash 073161/s61.9$0.30

110Grok 4.361/s53.3$2.50

110Llama 4 Maverick61/s34.4$1.00

110Mistral Large 361/s34.3$1.50

115Qwen3.6-27B60/s52.1$3.20

116Qwen3.5 397B59/s53.8$3.60

117Claude Fable 556/s74.0$50.00

117Grok 4.656/s67.8$6.00

117GPT-4.1 nano56/s26.3$0.40

117GPT-4o mini56/s26.2$0.60

121Qwen3.8-Flash-Next55/s64.3—

121GPT-5.1-Codex-Max55/s53.8$10.00

121GPT-5.1-Codex55/s52.9$10.00

124Grok 4.553/s63.7$6.00

125Hy452/s72.0$1.25

125GPT-5.552/s69.1$30.00

125GPT-5.3 Codex52/s62.0$14.00

128Grok 4.751/s69.1$6.00

128GPT-5.6 Luna51/s65.5$1.20

128Qwen3.7 Plus51/s57.7$1.28

131GPT-5.150/s55.3$10.00

132Claude Sonnet 449/s43.0$15.00

133Claude Opus 548/s73.8$25.00

133Kimi K2 Thinking48/s48.2$2.50

133Qwen3 235B 250748/s42.3$0.80

136Claude Opus 4.747/s62.0$25.00

136MiMo-V2.5-Pro47/s56.3$1.80

138GPT-6.1 Terra46/s77.1$5.00

139GPT-5.245/s59.0$14.00

139GLM-545/s51.6$3.20

139Claude Sonnet 4.545/s47.9$15.00

142Kimi K344/s66.9$16.00

142GPT-5.2-Codex44/s59.9$14.00

142Claude Sonnet 4.644/s57.0$15.00

142Qwen3.6 Plus44/s54.6$1.95

142Solar Pro 444/s52.1$0.60

142MiniMax M2.544/s50.2$1.08

142Qwen3 235B-A22B44/s38.1$1.82

149Kimi K3.1 Pro43/s75.7$10.00

149Qwen3.5-35B-A3B43/s49.4$1.00

151DeepSeek V4 Pro 081342/s65.5$4.44

152Qwen3.8 Max41/s65.6—

152Qwen3.8-Omni-Flash41/s62.7$0.47

152Phi-4-reasoning41/s27.1—

155Claude 4 Sonnet40/s39.5$15.00

156Claude Opus 4.539/s52.0$25.00

157GPT-5.438/s64.8$15.00

158Seed 2.1 Turbo37/s57.4$2.50

158Kimi K237/s39.6$2.30

158DeepSeek V337/s29.9$1.10

161Claude Opus 4.636/s58.0$25.00

161Kimi K2.536/s52.2$3.32

163Kimi K2.7 Code35/s57.7$3.50

163MiniMax M2.735/s51.2$0.84

163Gemma 4 31B35/s47.9Free

166GPT-6 Astra34/s78.0$50.00

166GPT-6.1 Sol34/s78.0$10.00

166MiniMax M2.134/s49.1$1.20

166GPT-OSS 120B34/s43.4$0.60

170Qwen3.8 Max (0902)33/s63.4$6.00

171Gemini 4 Carbon31/s82.7$10.00

171MiMo-V2.6-Pro31/s70.2$0.87

171Gemma 3 27B31/s25.3—

174GLM-5V-Turbo30/s48.1$4.00

174GLM-4.630/s41.8$2.00

176Claude 4.1 Opus Thinking29/s47.0—

176Llama 3.1 405B29/s27.0—

178Gemini 4 Argon28/s78.5$10.00

178MiMo-V2.6-Flash28/s64.3$0.28

178GLM-4.728/s50.0$2.20

178Mistral Small 3.128/s24.2—

182GLM-5-Turbo25/s51.9$4.00

182GLM-4.525/s45.7$2.20

184GPT-5.4 Pro21/s73.3$180.00

184Ternary Bonsai 2 27B21/s52.2$0.50

186DeepSeek-R1-052820/s44.6$2.15

186Qwen3.5-4B20/s44.3—

188GPT-5.5 Pro19/s74.8$180.00

188DeepSeek V3.219/s44.2$0.38

190DeepSeek-R118/s38.9$2.50

191Claude Opus 4.112/s47.5$75.00

192MiMo-V2.54/s54.8$2.00