Provider · Founded 2016 · Mountain View, USA

Groq: models, pricing and benchmarks

Inference company building LPU accelerators.

Models

in this index

Avg input

$2.30

per 1M tokens

Avg output

$10.23

per 1M tokens

Open weights

models

Best model

llama-3-3-groq

by average score

Website

Visit

https://groq.com

All Groq models

Sorted by average benchmark score. Prices are per million tokens and context is measured in tokens.

#	Model	Provider	Context	Input / 1M	Avg score
1	Llama 3.3 Groq	Groq	32768	$3.98	76.8
2	Llama 3.3 Groq Chat	Groq	32768	$4.06	76.8
3	Llama 3.3 Groq Instruct	Groq	65536	$4.08	76.7
4	Llama 3.3 Groq Base	Groq	16384	$3.24	73.5
5	Mixtral Groq	Groq	16384	$3.37	73.1
6	Mixtral Groq Instruct	Groq	16384	$3.25	72.5
7	Mixtral Groq Chat	Groq	8192	$3.09	72.0
8	Mixtral Groq Base	Groq	32768	$2.66	68.4
9	Llama 3.3 Groq 8B	Groq	16384	$1.17	59.5
10	Llama 3.3 Groq 7B	Groq	16384	$1.01	56.7
11	Mixtral Groq 8B	Groq	32768	$0.77	54.8
12	Mixtral Groq 7B	Groq	4096	$0.70	53.4
13	Llama 3.3 Groq 1.5B	Groq	4096	$0.46	50.0
14	Mixtral Groq 1.5B	Groq	4096	$0.34	46.1

Best Groq models

multimodal

Llama 3.3 Groq

Llama 3.3 Groq is a multimodal model from Groq. It offers a 32,768-token context window, costs $3.98 per million input tokens and $16.20 per million output tokens, and averages 76.8 across the index's benchmark suite.

$3.98 in · 32,768 context · 76.8 avg multimodal

Llama 3.3 Groq Chat

Llama 3.3 Groq Chat is a multimodal model from Groq. It offers a 32,768-token context window, costs $4.06 per million input tokens and $19.98 per million output tokens, and averages 76.8 across the index's benchmark suite.

$4.06 in · 32,768 context · 76.8 avg multimodal

Llama 3.3 Groq Instruct

Llama 3.3 Groq Instruct is a multimodal model from Groq. It offers a 65,536-token context window, costs $4.08 per million input tokens and $22.27 per million output tokens, and averages 76.7 across the index's benchmark suite.

$4.08 in · 65,536 context · 76.7 avg open-weight

Llama 3.3 Groq Base

Llama 3.3 Groq Base is a open-weight model from Groq. It offers a 16,384-token context window, costs $3.24 per million input tokens and $12.28 per million output tokens, and averages 73.5 across the index's benchmark suite.

$3.24 in · 16,384 context · 73.5 avg open-weight

Mixtral Groq

Mixtral Groq is a open-weight model from Groq. It offers a 16,384-token context window, costs $3.37 per million input tokens and $15.23 per million output tokens, and averages 73.1 across the index's benchmark suite.

$3.37 in · 16,384 context · 73.1 avg

How Groq fits

Groq has 14 models in the index, of which 14 are open weights. The average input price across the lineup is $2.30 per million tokens, and the average output price is $10.23 per million tokens. The strongest model by average benchmark score is llama-3-3-groq.

Use the table above to compare individual models, or open a model page for its full benchmark breakdown. If you are choosing between providers, start with the comparison pages and then check context limits and licence terms.

FAQ

How many Groq models are in the index?

Groq has 14 models in this index, of which 14 are open weights.

What is the cheapest Groq model?

Mixtral Groq 1.5B is the cheapest Groq model by input price at $0.34 per million tokens.

What is the best Groq model?

Llama 3.3 Groq has the highest average benchmark score in the Groq lineup at 76.8.

Are Groq models open weights?

14 of the 14 Groq models in this index are open weights. The rest are proprietary.

Where is Groq based?

Groq was founded in 2016 and is headquartered in Mountain View, USA.