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

Name: Artificial Analysis Output Speed
Creator: Artificial Analysis

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

Leaderboard

Scores from Artificial Analysis Output Speed, published Sep 26, 2026

Rank	Model	Score
1	Gemini 3.8 FlashGoogle · as “Gemini 3.8 Flash (high)”	328 tok/s
2	DeepSeek V4.1 FlashDeepSeek · as “DeepSeek V4.1 Flash (Reasoning, Max Effort)”	233 tok/s
3	DeepSeek V4 Flash VisionDeepSeek · as “DeepSeek V4 Flash Vision (Reasoning, Max Effort)”	218 tok/s
4	Muse Spark 1.3Meta · as “Muse Spark 1.3 (max)”	214 tok/s
5	GPT-6 LunaOpenAI · as “GPT-6 Luna (max)”	154 tok/s
6	GPT-5.3 CodexOpenAI · as “GPT-5.3 Codex (xhigh)”	119 tok/s
7	GPT-5.6 TerraOpenAI · as “GPT-5.6 Terra (max)”	100 tok/s
8	DeepSeek V4 Pro 0813DeepSeek · as “DeepSeek V4 Pro 0813 (Reasoning, Max Effort)”	94 tok/s
9	Claude Opus 5.5Anthropic · as “Claude Opus 5.5 (Adaptive Reasoning, Xhigh Effort, Default Fallback)”	90 tok/s
10	GPT-6 SolOpenAI · as “GPT-6 Sol (max)”	86 tok/s
11	GLM-5.3Z.ai · as “GLM-5.3 (low)”	86 tok/s
12	Claude Sonnet 5Anthropic · as “Claude Sonnet 5 (Adaptive Reasoning, Max Effort)”	84 tok/s
13	Grok 4.7SpaceXAI · as “Grok 4.7 (xhigh)”	78 tok/s
14	Grok 4.6SpaceXAI · as “Grok 4.6 (xhigh)”	72 tok/s
15	Step 5 PreviewStepFun · as “Step 5 Preview”	69 tok/s
16	Claude Fable 5.1Anthropic · as “Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback)”	69 tok/s
17	GPT-6 AstraOpenAI · as “GPT-6 Astra (max)”	63 tok/s
18	Qwen3.8-Flash-NextAlibaba · as “Qwen3.8-Flash-Next”	54 tok/s
19	GLM 5.3 FlashZ.ai · as “GLM 5.3 Flash”	49 tok/s
20	MiMo-V2.6-ProXiaomi · as “MiMo-V2.6-Pro”	44 tok/s
21	Qwen3.8 27BAlibaba · as “Qwen3.8 27B (xhigh)”	44 tok/s
22	Qwen3.8 Max (0902)Alibaba · as “Qwen3.8 Max (0902)”	39 tok/s
23	Qwen3.8 2.4T A95BAlibaba · as “Qwen3.8 2.4T A95B”	39 tok/s
24	Kimi K3Moonshot AI · as “Kimi K3 (low)”	38 tok/s