Outmake Models
Models

Qwen3.8 27B

Alibaba Qwen · August 2026

GetHugging Face

For coding, research and long tasks, and last month's most downloaded of Outmake's 24 GB Mac picks on Hugging Face.

Rating
4.1
Downloads
17M
Parameters
28B
Context
256K
Download
7.1 to 56 GB
Commercial use
Yes

Runs well on M4 · 16 GB

Outmake picks the UD-IQ2_S size: a 9.3 GB download that uses 10 GB of memory, and writes about 6.1 words (8.1 tokens) a second.

About

Designed to carry complex tasks with many steps through to the end. It reads pictures and video, and it can think before answering when that is switched on.

How it compares

Of the chat models Outmake recommends for a 24 GB Mac that LM Arena rates, its public vote ties it with Gemma 4 26B at the top.

In LM Arena's public vote for reading pictures, it ranks 3rd of the 15 open models rated.

Outmake estimates it is the slowest of the 8 chat models it recommends for a 24 GB Mac. Qwen3.6 35B writes about 5.8 times as fast.

Good to know

It follows the Qwen3.5 and Qwen3.6 series and is built on the design of Qwen3.5.

Sources: Qwen/Qwen3.8-27B model card, Outmake's speed estimates, LM Arena and Hugging Face. Every line comes from one of these.

Ratings

4.4 out of 5, from how people voted on its answers.

Text
1436
Coding
1498
Vision
1262

LM Arena ratings of October 28, 2026 (CC BY 4.0): people vote between two answers without knowing which model wrote which. The bar ends at 1488, the best any open model has. The Arena tests the full model; a small size on a Mac answers a little worse.

Good at

  • Long multi-step coding tasks
  • Science and maths diagrams
  • Understanding documents
  • Hour-long videos

Licence

Apache 2.0. Commercial use: yes. The maker's page has the full terms.

Qwen3.8 27B runs well from a Mac with an M1 and 16 GB of memory, at UD-IQ2_S, writing about 3.5 words (4.7 tokens) a second. 56 of 59 Mac configurations run it well.

Compare with

The nearest models in size, judged for the same Mac.

Sizes

SizeDownloadMemory at 8KMemory at 32KOpen
UD-IQ4_XSSmaller15 GB15 GB16 GBOpen
UD-Q4_K_XLBalanced18 GB18 GB19 GBOpen
UD-Q6_KBest23 GB22 GB23 GBOpen
All 25 sizes
SizeDownloadMemory at 8KMemory at 32KOpen
UD-IQ1_S7.1 GB7.2 GBestimated8.1 GBOpen
UD-IQ1_M7.7 GB7.7 GBestimated8.6 GBOpen
UD-IQ2_XXS8.2 GB8.3 GBestimated9.1 GBOpen
UD-IQ2_S9.3 GB9.4 GBestimated10 GBOpen
UD-Q2_K_XL11 GB11 GBestimated12 GBOpen
UD-IQ3_XXS12 GB12 GBestimated13 GBOpen
UD-IQ3_S13 GB13 GBestimated14 GBOpen
UD-Q3_K_XL14 GB14 GBestimated15 GBOpen
UD-Q4_K_S16 GB16 GBestimated17 GBOpen
Q4_017 GB17 GBestimated18 GBOpen
UD-Q4_K_M17 GB17 GBestimated18 GBOpen
Q4_118 GB18 GBestimated19 GBOpen
UD-Q5_K_S20 GB20 GBestimated20 GBOpen
UD-Q5_K_M21 GB21 GBestimated21 GBOpen
UD-Q5_K_XL22 GB22 GBestimated22 GBOpen
UD-Q6_K_M24 GB24 GBestimated24 GBOpen
UD-Q6_K_L25 GB25 GBestimated26 GBOpen
UD-Q6_K_XL26 GB26 GBestimated27 GBOpen
UD-Q8_K_L29 GB29 GBestimated29 GBOpen
Q8_030 GB30 GBestimated30 GBOpen
UD-Q8_K_XL32 GB32 GBestimated33 GBOpen
BF1656 GB56 GBestimated56 GBOpen

Memory is what llama.cpp's own calculator reported for each size when the catalog was built, with the app's settings. Smaller, Balanced and Best are the three sizes the store names.

Run It Anywhere

Qwen3.8 27B is an open model, so any app that runs GGUF files can run it. These take the UD-Q4_K_XL size from unsloth straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/unsloth/Qwen3.8-27B-GGUF:UD-Q4_K_XL
  • llama.cppMac, Windows, Linuxllama-server -hf unsloth/Qwen3.8-27B-GGUF:UD-Q4_K_XL
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/unsloth/Qwen3.8-27B-GGUF:UD-Q4_K_XL
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBNone worth runningneeds 11 GB
8 GBNone worth runningneeds 11 GB
10 GBNone worth runningneeds 11 GB
11 GBUD-Q2_K_XL11 GB11 GB
12 GBUD-Q2_K_XL11 GB11 GB
16 GBUD-IQ4_XS15 GB15 GB
20 GBUD-Q5_K_S20 GB20 GB
24 GBUD-Q6_K23 GB22 GB
32 GBQ8_030 GB30 GB
48 GBUD-Q8_K_XL32 GB32 GB

The largest size, of those that keep three bits a weight or more, whose memory at an 8K chat is under 92% of the card's; the rest is the desktop's. Estimated from the files, the same way as for a Mac; speed on a PC is not estimated yet. A size that doesn't fit the card still runs, slower, with part of it in ordinary memory.

Tests

ScienceGPQA Diamond
89.2%
Expert questionsHumanity's Last Exam
30.8%

GPQA Diamond: questions PhD scientists write for each other. Humanity's Last Exam: questions written to be beyond today's models. Scores are the ones filed on the model's Hugging Face page or the board's own table, read 2026-11-05; some come from the maker.

On every Mac
MacSizeFitMemoryWritesReads
M6 · 16 GBUD-IQ2_SRuns well10 GB8.7226
M6 · 24 GBUD-IQ4_XSRuns well15 GB5.7226
M6 · 32 GBUD-IQ4_XSRuns great15 GB5.7226
M5 · 16 GBUD-IQ2_SRuns well10 GB884
M5 · 24 GBUD-IQ4_XSRuns well15 GB5.284
M5 · 32 GBUD-IQ4_XSRuns great15 GB5.284
M5 Pro · 24 GBUD-IQ4_XSRuns well15 GB10167
M5 Pro · 48 GBUD-Q6_KRuns great23 GB7167
M5 Pro · 64 GBUD-Q6_KRuns great27 GB7209
M5 Max · 36 GBUD-IQ4_XSRuns great16 GB15334
M5 Max · 48 GBUD-Q6_KRuns great23 GB14418
M5 Max · 64 GBUD-Q6_KRuns great27 GB14418
M5 Max · 128 GBUD-Q6_KRuns great32 GB14418
M5 Ultra · 96 GBUD-Q6_KRuns great32 GB26668
M5 Ultra · 256 GBUD-Q6_KRuns great32 GB26835
M5 Ultra · 512 GBUD-Q6_KRuns great32 GB26835
M4 · 16 GBUD-IQ2_SRuns well10 GB6.128
M4 · 24 GBUD-IQ4_XSRuns well15 GB4.128
M4 · 32 GBUD-IQ4_XSRuns great15 GB4.128
M4 Pro · 24 GBUD-IQ4_XSRuns well15 GB8.556
M4 Pro · 48 GBUD-Q6_KRuns great23 GB6.156
M4 Pro · 64 GBUD-Q6_KRuns great27 GB6.156
M4 Max · 36 GBUD-IQ4_XSRuns great16 GB12112
M4 Max · 48 GBUD-Q6_KRuns great23 GB11140
M4 Max · 64 GBUD-Q6_KRuns great27 GB11140
M4 Max · 96 GBUD-Q6_KRuns great32 GB8.7112
M4 Max · 128 GBUD-Q6_KRuns great32 GB11140
M3 · 8 GBUD-IQ4_XSToo big25 GB
M3 · 16 GBUD-IQ2_SRuns well10 GB5.223
M3 · 24 GBUD-IQ4_XSRuns well15 GB3.423
M3 Pro · 18 GBUD-IQ3_XXSRuns well12 GB6.241
M3 Pro · 36 GBUD-IQ4_XSRuns great16 GB541
M3 Max · 36 GBUD-IQ4_XSRuns great16 GB9.587
M3 Max · 48 GBUD-Q6_KRuns great23 GB8.7116
M3 Max · 64 GBUD-Q6_KRuns great27 GB8.7116
M3 Max · 96 GBUD-Q6_KRuns great32 GB6.787
M3 Max · 128 GBUD-Q6_KRuns great32 GB8.7116
M3 Ultra · 96 GBUD-Q6_KRuns great32 GB16174
M3 Ultra · 256 GBUD-Q6_KRuns great32 GB16174
M3 Ultra · 512 GBUD-Q6_KRuns great32 GB16232
M2 · 8 GBUD-IQ4_XSToo big25 GB
M2 · 16 GBUD-IQ2_SRuns well10 GB5.223
M2 · 24 GBUD-IQ4_XSRuns well15 GB3.423
M2 Pro · 16 GBUD-IQ2_SRuns well10 GB9.846
M2 Pro · 32 GBUD-IQ4_XSRuns great15 GB6.646
M2 Max · 32 GBUD-IQ4_XSRuns great15 GB1286
M2 Max · 64 GBUD-Q6_KRuns great27 GB8.786
M2 Max · 96 GBUD-Q6_KRuns great32 GB8.7108
M2 Ultra · 64 GBUD-Q6_KRuns great27 GB16171
M2 Ultra · 128 GBUD-Q6_KRuns great32 GB16171
M2 Ultra · 192 GBUD-Q6_KRuns great32 GB16171
M1 · 8 GBUD-IQ4_XSToo big25 GB
M1 · 16 GBUD-IQ2_SRuns well10 GB3.518
M1 Pro · 16 GBUD-IQ2_SRuns well10 GB937
M1 Pro · 32 GBUD-IQ4_XSRuns great15 GB6.237
M1 Max · 32 GBUD-IQ4_XSRuns great15 GB1163
M1 Max · 64 GBUD-Q6_KRuns great27 GB8.163
M1 Ultra · 64 GBUD-Q6_KRuns great27 GB14127
M1 Ultra · 128 GBUD-Q6_KRuns great32 GB14127

Writes and reads are words a second (a token is about three quarters of a word), 4K tokens into a chat, estimated unless marked measured. The size is the one the store offers that Mac. Constants fitted October 11, 2026; how, and how wrong.

Badge for the model card

Shows on Qwen3.8 27B's Hugging Face card which Macs run it well, linked to this page. It updates every morning.

Runs on Mac badge

Sizes and files are read from unsloth/Qwen3.8-27B-GGUF on Hugging Face and checked every morning. The rating is the model's score on the LM Arena text leaderboard (CC BY 4.0), where people vote between two answers without knowing which model wrote which. Catalog of November 4, 2026.

Your Computer

Every model is judged for this Mac. Both are under About This Mac, in the Apple menu.

Every Mac · Every graphics card

Civitai

Open in OutmakeHugging Face

New to Outmake? Download the free beta