Skip to content
local-ai

Qwen3 14B

Alibaba · Text generation · 14B · 131k context · Released 29 April 2025

Commercial use permitted Open weights Runs on CPU Apple Silicon

The 14-billion-parameter Qwen3 model, a strong middle ground: noticeably more capable than the 8B while still fitting a 16GB card at a good quantisation. Apache 2.0, with an optional thinking mode and 128k context.

Strengths

  • A clear step up in reasoning and reliability over the 8B
  • Fits a 16GB card at Q8, or a 12GB card at Q4
  • Apache 2.0, long context, optional thinking mode

Weaknesses

  • Needs more memory than the popular 7B to 8B class
  • Thinking mode increases latency and token use
  • Still short of the 32B and 70B class on the hardest tasks

Hardware requirements

QuantisationApprox. VRAMNotes
Q4_K_M~9GBFits a 12GB card with short context
Q8_0~15.5GBNear-lossless, a good fit for 16GB cards
FP16~28GBFull precision, needs 32GB or more

Also runs on CPU (slower). Optimised builds available for Apple Silicon.

What you'd need to run this

Roughly what a machine to run this would need, at up to three levels of quality. Memory is the deciding factor.

Minimum to run it

Q4_K_M · ~9GB needed

One 16GB GPU

Intel Arc A770 16GB

or a Mac or mini-PC with unified memory, if you prefer no discrete GPU, Mac mini M5 Pro .

At least 32GB of system RAM alongside the card.

around £700–£1,100

What else 16GB runs →

For good quality

Q8_0 · ~15.5GB needed

One 24GB GPU

NVIDIA Tesla P40

or a Mac or mini-PC with unified memory, if you prefer no discrete GPU, Mac mini M5 Pro .

32–64GB of system RAM alongside the card.

around £700–£1,100

What else 24GB runs →

Best quality

FP16 · ~28GB needed

64GB of unified memory

Mac mini M4 Pro

or a single high-memory workstation card, NVIDIA RTX 6000 Ada Generation .

Unified memory is shared with the model, so it is already counted above.

around £1,100–£2,000

See the Affordable Mac build build What else 64GB runs →

Licence

Apache 2.0 — read the licence

Benchmarks

BenchmarkScoreSourceAs of
GPQA Diamond64.0 Qwen3 technical report (thinking mode) May 2025
AIME 202479.3 Qwen3 technical report (thinking mode) May 2025
AIME 202570.4 Qwen3 technical report (thinking mode) May 2025

How it compares

How this model’s reported scores sit against other models we cover, on the same benchmarks. This model is highlighted.

AIME 2025

higher is better

AIME 2024

higher is better

GPQA Diamond

higher is better

Qwen3 14B: common questions

What hardware do I need to run Qwen3 14B?
At its most compressed (Q4_K_M) it needs roughly 9GB of VRAM, and about 15.5GB for good quality. VRAM figures are approximate and depend on context length and settings.
Is Qwen3 14B free for commercial use?
Yes. Qwen3 14B is licensed under Apache 2.0, which permits commercial use with no meaningful conditions.
Can I run Qwen3 14B on Apple Silicon?
Yes. Qwen3 14B has builds optimised for Apple Silicon, through MLX or GGUF on a Mac.
Does Qwen3 14B run on CPU?
Yes, Qwen3 14B can run on the CPU, though generation is slower than on a GPU.
What is Qwen3 14B's context window?
Qwen3 14B has a context window of 131,072 tokens, about 131k.

Availability

Where to get quantised weights

Some of the best quantised weights are made by the community, not the model’s authors. Look this model up on these providers:

  • Unsloth GGUF (Dynamic 2.0, imatrix)

    Dynamic and imatrix GGUF quants that often hold quality better than a plain quant at the same bit-width, especially at 4-bit and below.

  • Bartowski GGUF Q2-Q8 (imatrix)

    A wide, reliable range of imatrix GGUF quants, typically Q2 through Q8.

  • MLX community MLX 4-bit and 8-bit

    MLX quants for Apple Silicon, usually 4-bit and 8-bit.

Recommended for

  • A capable general model on a 16GB card
  • Users who want more reliability than an 8B without needing a 24GB card

Run it with

Related guides

Glossary

Catalogue entry last verified 30 July 2026. Specifications change; verify anything you are about to spend money on.