Every AI leaderboard

Leaderboards28Models ranked986

28 leaderboards · 986 models ranked

Artificial Analysis Output Speed leaderboard: who is #1

Artificial Analysis · Output speed, higher is better · 43 models · published Nov 7, 2026

Artificial Analysis

Top 25 of 43 models, best variant per model as published

#ModelScoreCorrect
1New, entered at #1 since Nov 5, 2026MiMo-V2.6-FlashXiaomi
Unverified
64 tok/s
213up from #15 since Nov 5, 2026Qwen3.8 Max (0902)Alibaba
Unverified
60 tok/s
32up from #5 since Nov 7, 2026Qwen3.8 27Bas “Qwen3.8 27B (xhigh)”
Unverified
44 tok/s
4New, entered at #4 since Nov 5, 2026MiMo-V2.6-ProXiaomi
Unverified
43 tok/s
51up from #6 since Nov 7, 2026GLM 5.3 Flashas “GLM-5.3-Flash”
Unverified
39 tok/s
61up from #7 since Nov 7, 2026Kimi K3.1 ProMoonshot AI
Unverified
36 tok/s
74down from #3 since Nov 7, 2026Gemini 4 Carbonas “Gemini 4 Carbon (low)”
Unverified
35 tok/s
81up from #9 since Nov 5, 2026GLM-5.3as “GLM-5.3 (max)”
Unverified
32 tok/s
95down from #4 since Nov 5, 2026Claude Sonnet 5.5as “Claude Sonnet 5.5 (max)”
Unverified
31 tok/s
101up from #11 since Nov 4, 2026Gemini 4 Argonas “Gemini 4 Argon (high)”
Unverified
28 tok/s
111down from #10 since Nov 7, 2026Claude Opus 5.5as “Claude Opus 5.5 (max)”
Unverified
25 tok/s
121down from #11 since Nov 5, 2026Grok 4.7as “Grok 4.7 (xhigh)”
Unverified
24 tok/s
132up from #15 since Nov 7, 2026DeepSeek V4 Pro 0813as “DeepSeek V4 Pro 0813 (max)”
Unverified
23 tok/s
141down from #13 since Nov 7, 2026Claude Fable 5.1as “Claude Fable 5.1 (max)”
Unverified
22 tok/s
151down from #14 since Nov 7, 2026GPT-6.1 Solas “GPT-6.1 Sol (max)”
Unverified
22 tok/s
161up from #17 since Nov 7, 2026GPT-6 Astraas “GPT-6 Astra (max)”
Unverified
21 tok/s
171down from #16 since Nov 7, 2026MiniMax-M3.1-FlashMiniMax
Unverified
17 tok/s
18New, entered at #18 since Nov 7, 2026Inkling 2 SmallThinking Machines
Unverified
17 tok/s
1918down from #1 since Nov 5, 2026Claude Haiku 5.5as “Claude Haiku 5.5 (max)”
Unverified
16 tok/s
201up from #21 since Nov 7, 2026Gemini 3.8 Flashas “Gemini 3.8 Flash (high)”
Unverified
15 tok/s
21New, entered at #21 since Nov 7, 2026Mistral Large 4 Highas “Mistral Large 4 (high)”
Unverified
14 tok/s
223down from #19 since Nov 7, 2026Step 5 PreviewStepFun
Unverified
14 tok/s
23New, entered at #23 since Nov 7, 2026Tencent Hy4Tencent
Unverified
14 tok/s
2418down from #6 since Nov 5, 2026GPT-6.1 Terraas “GPT-6.1 Terra (max)”
Unverified
14 tok/s
253down from #22 since Nov 7, 2026DeepSeek V4.1 Flashas “DeepSeek V4.1 Flash 1017 (max)”
Unverified
14 tok/s
About this board and its history

What it measures

Median output tokens per second on the first-party API, measured over the past two weeks.

Rank at each month's end

12345678910SepOctNow1MiMo-V2.6-Flash2Qwen3.8 Max (0902)3Qwen3.8 27B4MiMo-V2.6-Pro5GLM 5.3 Flash6Kimi K3.1 Pro5212
MiMo-V2.6-FlashQwen3.8 Max (0902)Qwen3.8 27BMiMo-V2.6-ProGLM 5.3 FlashKimi K3.1 ProGemini 3.8 Flashdeepseek-v4-flashClaude Haiku 5.5DeepSeek V4.1 Flash

Source

Scores as published by Artificial Analysis on the capture date (2026-11-09). Source: artificialanalysis.ai models page data.

artificialanalysis.ai/leaderboards/models