Qwen-72B on Fireworks AI

Name: Qwen-72B on Fireworks AI
Brand: Alibaba
Price: 0.90 USD

Qwen · Alibaba

Provisioned

Last refreshed 2026-04-18. Next refresh: weekly.

Why use Qwen-72B on Fireworks AI?

Fireworks AI offers Qwen-72B with pay-as-you-go pricing at $0.90/1M input tokens. Fireworks AI offers a generative AI platform as a service, focusing on rapid product iteration and cost-efficient AI deployment.

Input / 1M

$0.90

Output / 1M

$0.90

Cache

Not sourced

Batch

Not sourced

Setup recipe

Python + curl

Install

pip install openai

Auth

export FIREWORKS_API_KEY=...

Call

import os
from openai import OpenAI
client = OpenAI(
    api_key=os.environ["FIREWORKS_API_KEY"],

Model ID

qwen-72b

Request example

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["FIREWORKS_API_KEY"],
    base_url="https://api.fireworks.ai/inference/v1"
)
response = client.chat.completions.create(
    model="qwen-72b",
    messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)

Gotchas

Fireworks model IDs use "accounts/fireworks/models/{model-name}" format, e.g. "accounts/fireworks/models/llama4-scout-instruct-basic" or "accounts/fireworks/models/deepseek-r1".
The examples expect FIREWORKS_API_KEY; rename it only if your application config maps the new variable.

Pricing

Type	Price (per 1M)
Input tokens	$0.90
Output tokens	$0.90

Capabilities

No model capability flags are currently sourced.

About Qwen-72B

The Qwen-72B is a potent large language model developed by Alibaba Cloud, featuring 72 billion parameters and built on the Transformer architecture. This model introduces enhancements such as SwiGLU activation, attention QKV bias, and group query attention for efficient performance on complex language tasks. Trained on a dataset of over 3 trillion tokens, it excels in multiple domains, including language understanding and generation, code generation, and translation, across multiple languages like Chinese and English. With an impressive context length of up to 32,000 tokens and a vocabulary exceeding 150,000 tokens, it manages extensive text inputs well. However, it bears limitations like inherited training data biases and lack of common sense reasoning. Despite these constraints, it remains a strong performer across various benchmarks, with a subsequent version, Qwen2-72B, offering further advancements.