Can I run Qwen2.5 72B locally?

Yes, the weights are open and available on Hugging Face under the Tongyi Qianwen licence. At 72B parameters you will need substantial VRAM — quantised GGUF versions reduce the hardware requirement considerably.

What does it cost and how much context does it support?

Via API it runs $0.36 per million input tokens and $0.40 per million output tokens. The model supports up to 128K tokens of context, though the default configuration is set to 32K and extending it requires enabling YaRN in the config.

How does the 72B differ from smaller Qwen2.5 models?

Qwen2.5 ranges from 0.5B to 72B; the 72B is the largest and most capable, with the best performance on coding, math and structured output tasks — the smaller variants trade capability for speed and lower hardware cost.

Qwen2.5 72B Instruct

2591°

#llm #qwen #coding #math #multilingual #local

Run this model

Offers / Ads

OpenRouter Terminal / API

Pay per token

Chat with this model in the browser or call it from your code with one API key.

qwen | 32,768 token context | $0.36/M input, $0.40/M output. Qwen2.5 72B is the latest series of Qwen large language models. Qwen2.5 brings the following improvements upon Qwen2: Significantly more knowledge and has greatly improved capabilities in coding and mathematics, thanks to our specialized expert models in these domains. Significant improvements in instruction following, generating long texts (over 8K tokens), understanding structured data (e.g, tables), and generating structured outputs especially JSON. More resilient to the diversity of system prompts, enhancing role-play…qwen | 32,768 token context | $0.36/M input, $0.40/M output. Qwen2.5 72B is the latest series of Qwen large language models. Qwen2.5 brings the following improvements upon Qwen2: Significantly more knowledge and has greatly improved capabilities in coding and mathematics, thanks to our specialized expert models in these domains. Significant improvements in instruction following, generating long texts (over 8K tokens), understanding structured data (e.g, tables), and generating structured outputs especially JSON. More resilient to the diversity of system prompts, enhancing role-play implementation and condition-setting for chatbots. Long-context Support up to 128K tokens and can generate up to 8K tokens. Multilingual support for over 29 languages, including Chinese, English, French, Spanish, Portuguese, German, Italian, Russian, Japanese, Korean, Vietnamese, Thai, Arabic, and more. Usage of this model is subject to Tongyi Qianwen LICENSE AGREEMENTOpens in new tab.

Guides for this model

Best local LLM by use case (October 2026): the right open model for every job

How much VRAM do I actually need for local AI?

What is a GGUF file? Quantized models explained (Q4, Q8, fp8, NF4)

All guides

Captain's notes

Written by Captain, the house editor

Notes from 2026-10-13

Qwen2.5 72B Instruct is Alibaba's flagship open-weights instruction model, sitting at the top of the Qwen2.5 family. It stands out for strong coding and math, reliable JSON output, and genuine multilingual reach across 29 languages. If you want a capable, self-hostable model that handles structured tasks without constant coaxing, this is a serious option.

Known for

Strong coding and mathematics, reinforced by specialized training data
Reliable structured output, especially JSON and table understanding
Multilingual across 29 languages including Arabic, Japanese and Korean
Handles long generations up to 8K tokens without trailing off
Resilient to varied system prompts; behaves well in roleplay and chatbot setups

Watch out for

Default config caps context at 32K; reaching 128K requires manual YaRN configuration.
The Tongyi Qianwen licence applies — not a standard open-source licence; check commercial terms.
72B means serious VRAM requirements for local use; plan for multi-GPU or quantised weights.

Best for

Prompts to try

Debug and explain code

Here is a Python function that is supposed to parse a CSV file and return a list of dicts, but it silently drops rows when a field contains a comma inside quotes. Find the bug, fix it, and explain in plain terms why the original failed.

Extract structured JSON

Read the following unstructured meeting notes and return a JSON object with keys: date, attendees (array), decisions (array), and action_items (array with owner and deadline for each). Notes: 'March 3rd catch-up. Sarah, Tom, and Priya. We agreed to launch the beta on March 20th. Tom will update the docs by March 10th. Priya handles user comms by March 15th.'

Multilingual translation check

Translate the following English product disclaimer into French, Spanish, and Arabic. Then review each translation for any term that might carry an unintended legal or cultural meaning in that market, and flag it with a brief note. Text: 'Results may vary. This product is not intended to diagnose, treat, cure, or prevent any disease.'

Compare with

Qwen3 30B A3B Thinking 2507
API-only model
81,920 token context · $0.20/M input, $2.40/M output
Qwen3.5-27B
Open-weights model
262,144 token context · $0.26/M input, $2.60/M output
Qwen2.5 Coder 32B Instruct
Open-weights model
32,768 token context · $0.66/M input, $1.00/M output
Qwen3 235B A22B Instruct 2507
Open-weights model
262,144 token context · $0.09/M input, $0.55/M output

More text models from qwen

Questions people ask

Can I run Qwen2.5 72B locally?: Yes, the weights are open and available on Hugging Face under the Tongyi Qianwen licence. At 72B parameters you will need substantial VRAM — quantised GGUF versions reduce the hardware requirement considerably.
What does it cost and how much context does it support?: Via API it runs $0.36 per million input tokens and $0.40 per million output tokens. The model supports up to 128K tokens of context, though the default configuration is set to 32K and extending it requires enabling YaRN in the config.
How does the 72B differ from smaller Qwen2.5 models?: Qwen2.5 ranges from 0.5B to 72B; the 72B is the largest and most capable, with the best performance on coding, math and structured output tasks — the smaller variants trade capability for speed and lower hardware cost.

Captain's verdict
A well-rounded open-weights workhorse that earns its size on structured tasks, code and multilingual work — just sort out the licence before shipping.

Model card on Hugging Face

Captain writes these notes from the model's own description and model card. Prices and context sizes are the provider's at the time of writing and can change.

Ranking badge

Updated Oct 22, 2026

Made this model? Show this badge on your page.

Paste it into your Hugging Face card or Civitai description.

Change colors / personalize / get code

Not this creator?

Get your own badge

How this score adds up

This model has no popularity data from any source yet, so it carries a stand-in number below 7,000° that keeps the lists complete. It never ranks above a model with data. The breakdown appears here as soon as a source reports it.

Glossary: the terms used here

Heat score: One number for how much a model is used and wanted right now, from 0° to 9,999°. Every list on this site that says hottest is sorted by it and nothing else.
Floor: The 7,000° every model with real data starts from. The scale runs from 0° to 9,999°; the part below the floor is reserved for stand-ins (models without data), so that a model with data always ranks above one without. The floor is the same for everyone and decides nothing; the degrees above it are the measurement.
Download-equivalents: Everything the score treats as a download, added up: downloads themselves, each torrent seeder as 1,000, each AI Horde image and each run on a generation site as one.
Log scale: Counting by tens, not by ones: going from 1,000 to 10,000 downloads adds the same degrees as going from 10,000 to 100,000. It keeps giants from crowding out everything else.
Likes: Thumbs up and likes from every source, worth half a download each on the log scale.
Adoption rate: How many downloads a day a model gets right now, from the difference between two readings of a source. A quarter of the weight of the download term.
Measured and estimated: Measured means two readings of a source exist and the difference is real. Estimated means only today's total is known and the past is spread over the model's life; the site says which one it is showing.
Seeder: Someone who keeps a copy of the weights online for others to download over a torrent, or a hosted mirror doing the same. Each one counts as 1,000 downloads.
AI Horde: A volunteer network that generates images for free on donated computers. The number of images it has made with a model is public, and each counts as one download.
Generation sites: Sites that host their own copies of a model and let people generate with it, such as PixAI, SeaArt and Yodayo. Their downloads count like Civitai's and each generation run as one download.
OpenRouter tokens: A text model served through an API, such as Claude or GPT, has no weights to download, so its usage is measured in tokens: the pieces of text it read and wrote through OpenRouter, a gateway many apps use, over the last 30 days. Every 10,000,000 tokens count as one download. Source: OpenRouter (openrouter.ai/rankings), CC BY 4.0.
Full model (checkpoint): The weights themselves, the file that generates on its own. It gets a bonus because add-ons depend on it.
LoRA and embedding: Small add-on files that steer someone else's weights toward a style, a character or a concept. They cannot run alone.
Family and weights: The base a model is built on, such as SD 1.5, SDXL, Flux or Wan. Every model inherits the release date of its family's weights.
Age: Degrees taken off for how long ago the family's weights came out: 150° a year, up to 750°. It keeps old downloads that had years to pile up from freezing the top of the list.
Family boost: A small, bounded nudge for LoRAs of a family this site is pushing forward right now, always listed in the breakdown.
Stand-in score: A stable number below the floor given to a model no source reports on yet, so lists stay complete. It never ranks above a model with data.
New! (grace period): A model less than 7 days old shows New! instead of a score and holds no place in the rankings. Its score is computed from day one, but a model one day old grows faster than anything, so without the wait the week's #1 would always be whatever came out yesterday rather than what held up for a week.
Trend: How a model's growth over a period compares with the typical model's growth over the same period. Faster than typical reads as a rise, slower as a fall.
Typical model: The middle of the catalog: half the models with a figure grew faster over the period, half slower.

Similar Models

AI Models for Image, Video & Text Generation

Qwen2.5 72B Instruct is a text model for the Text / LLM family listed on ModelTrends.ai, a read-only catalog of open source AI image, video and text models. Compare it with other Text / LLM models and check its heat score to see how it is trending.

More Text / LLM models llm qwen coding math multilingual local data extraction

Browse by Model Family