---
title: "Groq API pricing per token: 4 models | BroadRouter"
description: "Groq API pricing per 1M tokens for 4 models, from USD 0.075 input, with cached rates, cheapest-host checks and maker price comparison."
url: https://broadrouter.com/llm-api-pricing/providers/groq/
lang: en
updated: 2026-11-14
publisher: BroadRouter (a BroadNet site); API by BroadNet Inference, https://inference.broadnet.ai
---

# Groq API pricing

Updated 2026-11-14

Groq lists standard prices for 4 models. The cheapest is GPT OSS 20B at USD 0.075 input and USD 0.30 output per 1M tokens; the dearest is Qwen3.8 27B at USD 0.80 / USD 4. It publishes cached-input prices for 3 of them; batch, flex or priority pricing is published as a rule rather than per-model prices (see tiers below).

Groq hosts models it does not train. By maker: OpenAI 3, Qwen 1. 4 prices, each read from [Groq's own pricing page](https://console.groq.com/docs/models) or one of 3 linked model pages, checked 2026-11-14. BroadRouter is BroadNet's own site and does not sell Groq's services. Automated re-check 2026-11-16: 4 of 4 prices still matched the provider's page.

Models with a standard price

4

Median input / output

USD 0.1125 / USD 0.45

Cached-input prices

3 of 4

Cheapest host for

0 of 4 shared models

## Which models is Groq the cheapest provider for?

None. Every one of the 4 models that Groq shares with other providers is cheaper elsewhere at the standard tier. The closest is [GPT OSS Safeguard 20B](https://broadrouter.com/llm-api-pricing/gpt-oss-safeguard-20b/), where Groq ranks 2 of 3. It is the dearest provider for 2.

## How does Groq price the most widely hosted models?

The 4 models Groq sells that the most other providers also sell, with Groq's rank among them on the 3:1 blend.

Groq on the most widely hosted models
| Model | Groq | Rank | Cheapest |
| --- | --- | --- | --- |
| [OpenAI GPT OSS 120B](https://broadrouter.com/llm-api-pricing/gpt-oss-120b/) | USD 0.15 / USD 0.60 | 11 of 21 | CoreWeave Forge USD 0.03 / USD 0.17 |
| [Qwen3.8 27B](https://broadrouter.com/llm-api-pricing/qwen3.8-27b/) | USD 0.80 / USD 4 | 15 of 15 | Chutes USD 0.039 / USD 1.25 |
| [GPT OSS 20B](https://broadrouter.com/llm-api-pricing/gpt-oss-20b/) | USD 0.075 / USD 0.30 | 9 of 10 | AkashML USD 0.018 / USD 0.09 |
| [GPT OSS Safeguard 20B](https://broadrouter.com/llm-api-pricing/gpt-oss-safeguard-20b/) | USD 0.075 / USD 0.30 | 2 of 3 | Amazon Bedrock USD 0.07 / USD 0.20 |

## Which model makers does Groq sell?

Groq prices models from 2 makers. OpenAI leads with 3 models. Standard prices, USD per 1M input / output tokens.

Groq models by maker
| Maker | Models | Cheapest model | Median input / output |
| --- | --- | --- | --- |
| OpenAI | 3 | GPT OSS 20B USD 0.075 / USD 0.30 | USD 0.075 / USD 0.30 |
| Qwen | 1 | Qwen3.8 27B USD 0.80 / USD 4 | — |

## Groq price list

USD per 1M tokens: 4 prices for 4 models, one row per model with any other service tiers stacked under the standard price. Sources checked 2026-11-14 unless a row says otherwise. "vs maker" compares a standard price with the model maker's own API on the 3:1 blend; a minus means Groq is cheaper. Of 1 hosted model price with a maker price to compare, 0 are below the maker's and 1 above.

Groq API prices by model
| Model | Tier | Input | Cached | Output | Context | vs maker | Source |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [GPT OSS 20B](https://broadrouter.com/llm-api-pricing/gpt-oss-20b/) | standard | USD 0.075 | USD 0.037 | USD 0.30 | 131,072 | — | [source](https://console.groq.com/docs/model/openai/gpt-oss-20b) |
| [GPT OSS Safeguard 20B](https://broadrouter.com/llm-api-pricing/gpt-oss-safeguard-20b/) | standard | USD 0.075 | USD 0.037 | USD 0.30 | 131,072 | — | [source](https://console.groq.com/docs/model/openai/gpt-oss-safeguard-20b) |
| [OpenAI GPT OSS 120B](https://broadrouter.com/llm-api-pricing/gpt-oss-120b/) | standard | USD 0.15 | USD 0.075 | USD 0.60 | 131,072 | — | [source](https://console.groq.com/docs/model/openai/gpt-oss-120b) |
| [Qwen3.8 27B](https://broadrouter.com/llm-api-pricing/qwen3.8-27b/) | standard | USD 0.80 | — | USD 4 | 131,072 | +42% | [source](https://console.groq.com/docs/models) |

-   **Cached "—"** means Groq's page shows no cache price for that row; budget cached tokens at the input price.

[Report a wrong price or number](mailto:info@broadnet.ai?subject=Wrong%20price%20or%20number%20on%20Groq%20API%20pricing&body=Page%3A%20https%3A%2F%2Fbroadrouter.com%2Fllm-api-pricing%2Fproviders%2Fgroq%2F%0A%0AWhat%20looks%20wrong%20(model%2C%20figure%20shown%2C%20correct%20figure%20and%20a%20source%20link%20if%20you%20have%20one)%3A%0A) (email to info@broadnet.ai, with this page's link filled in).

## What service tiers does Groq publish?

Groq publishes no per-model batch, flex or priority prices, but states these rules: Batch API: "50% cost discount compared to synchronous APIs" (console.groq.com/docs/batch). Flex processing: "Pricing matches the on-demand tier", for paid customers only (console.groq.com/docs/flex-processing). Checked 2026-11-15.

Cached input: 3 of 4 standard prices include a cache-read price, a median 49% of the input price. [What cached input means](https://broadrouter.com/glossary/cached-input-tokens/).

**What we recorded about Groq's pricing page:** groq.com/pricing now redirects to the homepage; prices come from the GroqCloud Supported Models docs and each model page. Llama 3.1 8B, Llama 3.3 70B, Muse Spark Open and MiniMax M2.7 are listed as "Contact Sales" (no public token price). Prompt Guard classifiers, Whisper and Orpheus TTS excluded (not chat LLMs). Re-checked 2026-11-15: Llama 3.3 70B (llama-3.3-70b-versatile) is still "Contact Sales" on console.groq.com/docs/models and its model page says "This model is available to Enterprise customers."; no public token price, so no row.

## Is Groq cheaper than the model makers' own APIs?

No. Of 1 standard price for models whose maker sells its own API, 0 are below the maker's price, 1 above and 0 the same, on the 3:1 blend. The median difference is 42% above the maker. The biggest markup is Qwen3.8 27B, 42% above Alibaba Cloud.

## BroadNet Inference and Groq

BroadRouter is BroadNet's own site, so BroadNet is never ranked on this page. BroadNet Inference also serves 1 model that Groq prices: [Qwen3.8 27B](https://broadrouter.com/models/qwen3.8-27b/) at USD 0.175 / USD 2.17 against Groq's USD 0.80 / USD 4 (BroadNet is cheaper on the 3:1 blend). BroadNet prices are from its rate card, synced 2026-11-14. [BroadNet Inference pricing](https://broadrouter.com/pricing/).

## More on Groq

-   [Groq's own pricing page](https://console.groq.com/docs/models) (the source; checked 2026-11-14)
-   [Groq alternative: fast inference options compared](https://broadrouter.com/alternatives/groq-alternative/)
-   [LLM API free tiers and rate limits](https://broadrouter.com/data/llm-api-free-tiers-and-rate-limits/): Groq's free tier and default limits next to other providers'
-   [All 41 providers compared](https://broadrouter.com/llm-api-pricing/providers/) · [Cheapest LLM APIs right now](https://broadrouter.com/llm-api-pricing/cheapest/) · [Download every price (CSV, JSON)](https://broadrouter.com/data/llm-api-provider-prices/)

## Quick answers

### Is Groq OpenAI-compatible?

Yes. Groq's OpenAI compatibility docs cover chat completions and the Responses API, with notes on audio transcription. logprobs, logit\_bias, top\_logprobs and messages\[\].name return 400, and n must be 1 ([Groq, OpenAI compatibility](https://console.groq.com/docs/openai), checked 2026-11-15).

### What is Groq's API base URL?

`https://api.groq.com/openai/v1`, as Groq documents it ([Groq, OpenAI compatibility](https://console.groq.com/docs/openai), checked 2026-11-15).

### Does Groq keep API prompts and outputs?

Not by default. Groq says: "By default, Groq does not retain customer data for inference requests" ([Groq, your data](https://console.groq.com/docs/your-data), checked 2026-11-14).

### Does Groq train on API data?

No, unless you allow it. Groq's Services Agreement says Groq "is not permitted to use Inputs or Outputs for training or fine-tuning any AI Model Services or other models, unless explicitly granted permission or instructed by Customer" ([Groq Services Agreement](https://console.groq.com/docs/legal/services-agreement), checked 2026-11-14).

### Does Groq have a free tier?

Free plan with per-model limits; the Developer plan raises them. Default limits (Free plan, openai/gpt-oss-120b): 30 requests a minute, 8,000 tokens a minute and 1,000 requests a day. Free plan, also 200,000 tokens a day. Whisper models: 20 RPM and 2,000 requests a day ([Groq rate limits docs](https://console.groq.com/docs/rate-limits), checked 2026-11-14).

### Does Groq have a batch API?

Yes. Groq's Batch API docs state a "50% cost discount compared to synchronous APIs" ([Groq Batch API docs](https://console.groq.com/docs/batch), checked 2026-11-15).

## Questions

### How much does the Groq API cost per token?

Groq's standard prices run from USD 0.075 input and USD 0.30 output per 1M tokens (GPT OSS 20B) to USD 0.80 / USD 4 (Qwen3.8 27B). The median across 4 standard prices is USD 0.1125 input and USD 0.45 output per 1M, which is USD 0.0000001125 per input token. Input, cached input and output are billed separately.

### What is the cheapest model on Groq?

GPT OSS 20B, at USD 0.075 input and USD 0.30 output per 1M tokens (USD 0.1312 on a 3:1 input-to-output blend). The cheapest from other makers: Qwen3.8 27B (Qwen) at USD 0.80 / USD 4.

### Is Groq cheaper than other providers?

It depends on the model. Of 4 models that other providers also price, Groq is the cheapest standard price for none; another provider is cheaper for 4; and it is the dearest for 2. Its median rank is 10 of a median 12.5 providers, on a 3:1 input-to-output blend.

### Does Groq have batch, cached or priority pricing?

Cached input: yes, for 3 of 4 standard prices, at a median 49% of the input price. No batch, flex or priority prices are listed per model, but Groq states: Batch API: "50% cost discount compared to synchronous APIs" (console.groq.com/docs/batch). Flex processing: "Pricing matches the on-demand tier", for paid customers only (console.groq.com/docs/flex-processing). Checked 2026-11-15.

### How much does a chatbot cost per month on Groq?

For 2,000 requests a day, 1,500 input tokens (60% cached), 300 output, over 30 days: USD 10.10 on GPT OSS 20B, USD 10.10 on GPT OSS Safeguard 20B (the middle of Groq's list) and USD 144.00 on Qwen3.8 27B. [Try your own numbers](https://broadrouter.com/tools/llm-api-cost-calculator/).

Try it on your real workload. New accounts start with $5 of credit for Gemma and transcription. DeepSeek, GLM and Qwen3.8 need a top-up.

[Get started](https://inference.broadnet.ai/account/?utm_source=broadrouter&utm_medium=referral&utm_campaign=llm-api-pricing-providers-groq&utm_content=page-cta) ·

**Sources**

-   [Groq: pricing page (checked 2026-11-14)](https://console.groq.com/docs/models)
-   [Groq: GPT OSS 20B price source (checked 2026-11-14)](https://console.groq.com/docs/model/openai/gpt-oss-20b)
-   [Groq: GPT OSS Safeguard 20B price source (checked 2026-11-14)](https://console.groq.com/docs/model/openai/gpt-oss-safeguard-20b)
-   [Groq: OpenAI GPT OSS 120B price source (checked 2026-11-14)](https://console.groq.com/docs/model/openai/gpt-oss-120b)
-   [Groq: rate limits and free tier (checked 2026-11-14)](https://console.groq.com/docs/rate-limits)
-   [Groq, OpenAI compatibility (checked 2026-11-15)](https://console.groq.com/docs/openai)
-   [Groq, your data (checked 2026-11-14)](https://console.groq.com/docs/your-data)
-   [Groq Services Agreement (checked 2026-11-14)](https://console.groq.com/docs/legal/services-agreement)
-   [Groq Batch API docs (checked 2026-11-15)](https://console.groq.com/docs/batch)
