Every AI leaderboard

Leaderboards28Models ranked992

28 leaderboards · 992 models ranked

Artificial Analysis Output Speed what it measures, how it is scored and who leads it

Speed, price and usage · Artificial Analysis · Tokens per second

Leaderboard

Scores from Artificial Analysis Output Speed, published Nov 11, 2026

RankModelScore
1MiMo-V2.6-FlashXiaomi · as “MiMo-V2.6-Flash”
Unverified
63 tok/s
2Qwen3.8 Max (0902)Alibaba · as “Qwen3.8 Max (0902)”
Unverified
56 tok/s
3GLM 5.3 FlashZ.ai · as “GLM-5.3-Flash”
Unverified
45 tok/s
4Claude Fable 5.2Anthropic · as “Claude Fable 5.2 (max)”
Unverified
44 tok/s
5Qwen3.8 27BAlibaba · as “Qwen3.8 27B (xhigh)”
Unverified
40 tok/s
6Kimi K3.1 ProMoonshot AI · as “Kimi K3.1 Pro”
Unverified
37 tok/s
7MiMo-V2.6-ProXiaomi · as “MiMo-V2.6-Pro”
Unverified
36 tok/s
8Claude Sonnet 5.5Anthropic · as “Claude Sonnet 5.5 (max)”
Unverified
31 tok/s
9GLM-5.3Z.ai · as “GLM-5.3 (max)”
Unverified
29 tok/s
10Gemini 4 ArgonGoogle · as “Gemini 4 Argon (high)”
Unverified
29 tok/s
11Grok 4.7SpaceXAI · as “Grok 4.7 (xhigh)”
Unverified
27 tok/s
12Gemini 4 CarbonGoogle · as “Gemini 4 Carbon (high)”
Unverified
25 tok/s
13Claude Opus 5.5Anthropic · as “Claude Opus 5.5 (max)”
Unverified
25 tok/s
14Agnes 2.5 Pro ThetaSapiens Ai · as “Agnes 2.5 Pro Theta”
Unverified
24 tok/s
15DeepSeek V4 Pro 0813DeepSeek · as “DeepSeek V4 Pro 0813 (max)”
Unverified
24 tok/s
16GPT-6 AstraOpenAI · as “GPT-6 Astra (max)”
Unverified
23 tok/s
17GPT-6.1 SolOpenAI · as “GPT-6.1 Sol (max)”
Unverified
20 tok/s
18Inkling 2 SmallThinking Machines · as “Inkling 2 Small”
Unverified
20 tok/s
19DeepSeek V4.1 FlashDeepSeek · as “DeepSeek V4.1 Flash 1017 (max)”
Unverified
17 tok/s
20Claude Haiku 5.5Anthropic · as “Claude Haiku 5.5 (max)”
Unverified
15 tok/s
21MiniMax-M3.1-FlashMiniMax · as “MiniMax-M3.1-Flash”
Unverified
14 tok/s
22Gemini 3.8 FlashGoogle · as “Gemini 3.8 Flash (high)”
Unverified
14 tok/s
23GPT-6.1 TerraOpenAI · as “GPT-6.1 Terra (max)”
Unverified
14 tok/s
24Apodex 1.1Apodex · as “Apodex 1.1”
Unverified
14 tok/s
25Muse WonderMeta · as “Muse Wonder (max)”
Unverified
13 tok/s

About this benchmark

Task
Measure how fast a model streams tokens on its first-party API.
Dataset
Repeated live API calls over the past two weeks.
Method
Median output tokens per second after the first token.
Metric
Tokens per second
Organization
Artificial Analysis
Official leaderboard
artificialanalysis.ai/leaderboards/models

Sources: each benchmark's paper (arXiv) and its official site or leaderboard. Scores are the boards captured here every day, exactly as published.