Fastest AI Models -- Speed Rankings

When response time matters. Real-time applications, chatbots, and interactive tools need speed above all. Ranked by tokens per second output speed.

#		Model	Provider	Speed	Latency	Intelligence	Price/1M
1		NVIDIA: Nemotron 3 Super	NVIDIA	402 tok/s	0.59s	36.0	$0.10
2		Google: Gemini 2.5 Flash	Google	343 tok/s	0.31s	19.4	$0.30
3	A	Amazon: Nova Micro 1.0	Amazon	329 tok/s	0.34s	10.3	$0.04
4		OpenAI: gpt-oss-20b	OpenAI	304 tok/s	0.44s	24.5	$0.03
5		Mistral: Ministral 3 3B 2512	Mistral AI	291 tok/s	0.25s	11.2	$0.10
6		OpenAI: gpt-oss-120b	OpenAI	289 tok/s	0.49s	33.3	$0.04
7		OpenAI: GPT-5	OpenAI	221 tok/s	2.15s	44.4	$1.25
8	A	Amazon: Nova Lite 1.0	Amazon	219 tok/s	0.38s	12.7	$0.06
9		Mistral: Mistral Small 3	Mistral AI	216 tok/s	0.29s	15.1	$0.05
10		Google: Gemini 2.5 Flash Lite Preview 09-2025	Google	214 tok/s	5.29s	33.5	$0.10
11		Google: Gemini 3.1 Flash Lite Preview	Google	214 tok/s	5.29s	33.5	$0.25
12		Mistral: Devstral Small 1.1	Mistral AI	204 tok/s	0.34s	19.5	$0.10
13		xAI: Grok 3 Mini	xAI	199 tok/s	0.37s	32.1	$0.30
14		Meta: Llama 3.1 8B Instruct	Meta	197 tok/s	0.44s	11.8	$0.02
15		Meta: Llama 3 8B Instruct	Meta	197 tok/s	0.44s	11.8	$0.03
16		Google: Gemini 3 Flash Preview	Google	186 tok/s	0.70s	35.0	$0.50
17	A	Amazon: Nova 2 Lite	Amazon	176 tok/s	0.53s	18.0	$0.30
18		xAI: Grok Code Fast 1	xAI	172 tok/s	2.93s	28.7	$0.20
19		OpenAI: GPT-5 Codex	OpenAI	170 tok/s	4.79s	44.6	$1.25
20		Mistral: Mistral Small 3.1 24B	Mistral AI	159 tok/s	0.40s	12.7	$0.03
21		Mistral: Mistral Small 3.2 24B	Mistral AI	159 tok/s	0.40s	12.7	$0.07
22		OpenAI: GPT-4o Audio	OpenAI	157 tok/s	0.48s	17.3	$2.50
23		Mistral: Mistral 7B Instruct v0.1	Mistral AI	157 tok/s	0.29s	7.4	$0.11
24		Mistral: Mistral Small Creative	Mistral AI	157 tok/s	0.40s	10.2	$0.10
25		OpenAI: GPT-5 Nano	OpenAI	154 tok/s	0.68s	13.8	$0.05
26	A	Amazon: Nova Pro 1.0	Amazon	151 tok/s	8.94s	35.7	$0.80
27		OpenAI: o3	OpenAI	143 tok/s	6.39s	25.9	$2.00
28		Mistral: Devstral Medium	Mistral AI	143 tok/s	0.40s	18.7	$0.40
29		Mistral: Ministral 3 8B 2512	Mistral AI	141 tok/s	0.27s	14.8	$0.15
30		OpenAI: o4 Mini Deep Research	OpenAI	140 tok/s	17.07s	33.1	$2.00
31		NVIDIA: Nemotron Nano 12B 2 VL	NVIDIA	138 tok/s	0.55s	10.1	$0.20
32		xAI: Grok 4 Fast	xAI	137 tok/s	3.83s	35.1	$0.20
33		Meta: Llama 4 Maverick	Meta	129 tok/s	0.47s	18.4	$0.15
34		Meta: Llama 4 Scout	Meta	129 tok/s	0.45s	13.5	$0.08
35		Mistral: Ministral 3 14B 2512	Mistral AI	124 tok/s	0.29s	16.0	$0.20
36		NVIDIA: Nemotron Nano 9B V2	NVIDIA	123 tok/s	0.22s	14.8	$0.04
37		Google: Gemini 2.5 Pro Preview 06-05	Google	120 tok/s	21.69s	34.6	$1.25
38		Google: Gemini 2.5 Pro	Google	120 tok/s	21.69s	34.6	$1.25
39		Google: Nano Banana Pro (Gemini 3 Pro Image Preview)	Google	119 tok/s	21.66s	48.4	$2.00
40		Anthropic: Claude 3 Haiku	Anthropic	118 tok/s	9.88s	37.1	$0.25
41		Anthropic: Claude Haiku 4.5	Anthropic	118 tok/s	9.88s	37.1	$1.00
42		Google: Gemini 3.1 Pro Preview	Google	117 tok/s	21.91s	57.2	$2.00
43		Google: Gemini 3.1 Pro Preview Custom Tools	Google	117 tok/s	21.91s	57.2	$2.00
44		Google: Gemini 3 Pro Preview	Google	116 tok/s	3.88s	41.3	$2.00
45		OpenAI: GPT-5 Chat	OpenAI	116 tok/s	0.63s	21.8	$1.25
46		NVIDIA: Nemotron 3 Nano 30B A3B	NVIDIA	116 tok/s	1.38s	24.3	$0.05
47		OpenAI: o1-pro	OpenAI	115 tok/s	17.50s	30.8	$150
48	P	Perplexity: Sonar Pro Search	Perplexity	113 tok/s	1.01s	15.5	$3.00
49		OpenAI: GPT-3.5 Turbo (older v0613)	OpenAI	105 tok/s	0.44s	9.0	$1.00
50		OpenAI: GPT-3.5 Turbo 16k	OpenAI	105 tok/s	0.44s	9.0	$3.00

Benchmark data by Artificial Analysis