Groq model pricing.

prices as of · 11 models · 11 prices

Serves open-weight models on its own LPU hardware with a free tier.

Groq is a host selling 11 models, from $0.23 per million tokens for Llama 3.1 8B Instruct, read on 16 Sept 2026.

Groq prices

Kind of provider

Host

based in US

Where it runs

not stated

no regions published

How you pay

not stated

no payment terms published

Free tier

none published

nothing given away

Where Groq's prices land, and why

Groq publishes 11 prices we can compare, and 2 of them are below the list price for that model. 1 of those have a published reason behind them, so they rank like any other price. Nobody here has read its terms yet, so treat what it says about itself as its own claim.

Groq does not say whether it trains on what you send it. It does not say how long it keeps requests.

not yet checked: terms · privacy policy ↗ · status page ↗

Models to start with

all 11 prices ↓

Everything Groq sells, against the list price

Cheapest first inside each group. The share is this row against the price the lab that trained the model charges.

Cheaper, and we know why

1 price

Below list, with a published reason: hosting the open weights, a regional price list, credits, a stated discount.

  • Qwen3 32B

    Alibaba Cloud · 131k context

    Hosts the open weights on its own servers, so the price is its own, not a resale. Precision not disclosed.

    $0.29 / $0.59in / out per M

    0.30×of list price

Same price as the maker

10 prices

The same price the maker charges, sold by someone else.

  • Llama 3.1 8B Instruct

    Meta · 131k context

    The same price OpenRouter charges.

    $0.05 / $0.08in / out per M

    1.00×of list price

  • gpt-oss-20b

    OpenAI · 131k context

    150% above OpenRouter's price.

    $0.07 / $0.30in / out per M

    2.39×of list price

  • gpt-oss-safeguard-20b

    OpenAI · 131k context

    The same price OpenRouter charges.

    $0.07 / $0.30in / out per M

    1.00×of list price

  • Llama Guard 4 12B

    Meta · 164k context

    11% above OpenRouter's price.

    $0.20 / $0.20in / out per M

    1.11×of list price

  • gpt-oss-120b

    OpenAI · 131k context

    305% above OpenRouter's price.

    $0.15 / $0.60in / out per M

    3.74×of list price

  • Llama-3.3-70B-Instruct

    Meta · 128k context

    490% above OpenRouter's price.

    $0.59 / $0.79in / out per M

    4.13×of list price

  • MiniMax-M2.7

    MiniMax · 205k context

    100% above MiniMax's price.

    $0.60 / $1.80in / out per M

    1.71×of list price

  • Qwen3.6 27B

    Alibaba Cloud · 262k context

    The same price Alibaba Cloud charges.

    $0.60 / $3in / out per M

    0.89×of list price

  • Kimi K2 0905

    Moonshot AI · 262k context

    67% above Moonshot AI's price.

    $1 / $3in / out per M

    1.40×of list price

  • Qwen3.8 27B

    Alibaba Cloud · 1M context

    88% above Alibaba Cloud's price.

    $0.80 / $4in / out per M

    1.67×of list price

Frequently asked

What is Groq?
Serves open-weight models on its own LPU hardware with a free tier. It sells 11 of the models we track, and trades as Groq, Inc., based in US.
Is Groq cheaper than buying from the maker?
2 of its 11 prices sit below the list price for the model, and 1 of those have a published reason behind them, so they can lead a ranking.
Does Groq train on what you send it?
Groq does not say whether it trains on what you send it. It does not say how long it keeps requests.
How current are these Groq prices?
They come from the public catalogs we read, last read 16 Sept 2026. Prices exclude tax.