Can I run Qwen3-Coder-30B-A3B-Instruct locally?

Yes, the weights are open and available on Hugging Face at Qwen/Qwen3-Coder-30B-A3B-Instruct. Because only 3.3B parameters are active per forward pass, memory requirements are lower than a dense 30B model, though you should check the VRAM guide for your specific quantisation.

What does it cost via API, and how long is the context?

The API price is $0.07 per million input tokens and $0.28 per million output tokens. The native context window is 262,144 tokens, extendable to 1M tokens using Yarn if your inference setup supports it.

Does this model have a thinking or reasoning mode?

No. Qwen3-Coder-30B-A3B-Instruct supports only non-thinking mode and never produces chain-of-thought blocks in its output. If you need reasoning traces, you would need a different model from the Qwen3 family.

Qwen3 Coder 30B A3B Instruct

4688°

#llm #qwen #coding #agents #long context #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 | 262,144 token context | $0.07/M input, $0.28/M output. Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, and agentic tool use. Built on the Qwen3 architecture, it supports a native context length of 256K tokens (extendable to 1M with Yarn) and performs strongly in tasks involving function calls, browser use, and structured code completion. This model is optimized for instruction-following without “thinking mode”, and integrates well with OpenAI-compatible tool-use formats.

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

Qwen3-Coder-30B-A3B-Instruct is a Mixture-of-Experts coding model from Qwen with 30.5B total parameters but only 3.3B active per forward pass, making it punchy relative to its weight class. It targets agentic coding workflows — think CLINE, browser automation, multi-step tool use — with a native 256K-token context that covers most real repositories. Open weights and a low API price make it worth serious consideration if coding agents are your main workload.

Known for

Agentic coding and browser-use tasks with structured function-call support
256K native context, extendable to 1M tokens via Yarn, for repo-scale work
MoE efficiency: 30.5B parameters, only 3.3B active per inference pass
OpenAI-compatible tool-use format; integrates with CLINE and Qwen Code out of the box
Instruction-following only — no thinking mode, no chain-of-thought overhead

Watch out for

No thinking mode at all; if you need step-by-step reasoning traces, look elsewhere in the Qwen3 family.
1M-token extension requires Yarn configuration — not available by default in every inference stack.
Licence terms are not detailed here; check the Hugging Face repository before any commercial deployment.

Best for

Prompts to try

Refactor a messy function

Here is a Python function that reads a CSV file, filters rows where the 'status' column equals 'active', computes the mean of the 'score' column for those rows, and writes the result to a new CSV. It works but is hard to read and has no error handling. Refactor it for clarity and robustness, and add type hints throughout. [paste your function here]

Build a multi-step agent tool

Write a Python function called search_and_summarize that takes a search query string, calls a hypothetical search_web(query) tool returning a list of {title, url, snippet} dicts, picks the three most relevant results, fetches each page with fetch_page(url) returning plain text, and returns a single concise summary paragraph. Use OpenAI-compatible tool-call format and handle failures gracefully.

Explain a large codebase file

I am going to paste a long source file from an unfamiliar repository. Read it carefully and give me: (1) what this file's responsibility is within the broader system, (2) the three most important functions or classes and what each does, (3) any obvious bugs or code smells, and (4) the questions I should ask to understand the parts that are unclear. [paste your file here]

Compare with

Qwen3.7 Flash
API-only model
1,000,000 token context · $0.03/M input, $0.13/M output
Qwen3.8 Max (0902)
API-only model
1,000,000 token context · $2.00/M input, $6.00/M output
Qwen3.8 Omni Flash
API-only model
1,000,000 token context · $0.15/M input, $0.47/M output
Qwen3.7 Plus
API-only model
1,000,000 token context · $0.32/M input, $1.28/M output

More text models from qwen

Questions people ask

Can I run Qwen3-Coder-30B-A3B-Instruct locally?: Yes, the weights are open and available on Hugging Face at Qwen/Qwen3-Coder-30B-A3B-Instruct. Because only 3.3B parameters are active per forward pass, memory requirements are lower than a dense 30B model, though you should check the VRAM guide for your specific quantisation.
What does it cost via API, and how long is the context?: The API price is $0.07 per million input tokens and $0.28 per million output tokens. The native context window is 262,144 tokens, extendable to 1M tokens using Yarn if your inference setup supports it.
Does this model have a thinking or reasoning mode?: No. Qwen3-Coder-30B-A3B-Instruct supports only non-thinking mode and never produces chain-of-thought blocks in its output. If you need reasoning traces, you would need a different model from the Qwen3 family.

Captain's verdict
A genuinely cheap, locally runnable coding agent that punches above its active-parameter count — the MoE architecture is the whole point here.

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

Qwen3 Coder 30B A3B 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 agents long context local budget

Browse by Model Family