Model Follower.

Up to date

Rebuilt
2026-10-25 06:54 UTC
Release radar
117 on watch
Hub snapshot
2026-10-25
Re-checked
daily · every number sourced and dated

Models › Qwen2.5 family

Qwen2.5-Coder-32B

Hugging Face ↗ Follow changes

Qwen Apache 2.0 out 2 years ago

32.8B parameters, text in, text out. 1 provider serves it.

Verified 2026-10-25 Rank #730 in LLMs · score 71.7 method v1 Sources 3 Downloads 30d 2k · #577 of 619 tracked

Changed this week: 1 new quantized build on the Hub (24 in all).2026-10-24

Can I run it?20GB at INT4Estimatefits a 24 GB cardmethod
Can I use it?Apache 2.0Permissive; patent grantlicence file
Do people use it?2kdownloads, 30 days#577 of 619 we track · below the median+2 likes, downloads +14% this week171 likesHugging Face

Worth reading

Picked by rule from 26 tracked pieces: the lab's own post, the most shared, a field report, the latest.

Most discussed · 164 HN pts

Qwen2.5-Coder-32B is an LLM that can code well that runs on my Mac

Alibaba released the Qwen2.5-Coder-32B-Instruct model, an Apache 2.0 licensed open-source LLM.

  • The model requires a workaround for an SSL bug when using the standard library instead of requests or httpx.
  • Running the Ollama version was too slow because the author did not have the tool compiled to use the Mac's GPU.
simonwillison.net 2024-11-13 164 HN pts · 65 comments 635 words

25 best of 26 tracked

The 3 picks above, plus these 22. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

  1. Codekev: Jev-like model built on Qwen2.5-0.5Bgithub.com · 2026-09-19
  2. ArticleQwen2.5-VL-32B: Smarter and Lighterqwenlm.github.io · 2025-03-24 · official
  3. ArticleQwen2.5-1M: Deploy your own Qwen with context length up to 1M tokensqwenlm.github.io · 2025-01-26 · official
  4. ArticleQwen2.5: A Party of Foundation Modelsqwenlm.github.io · 2024-09-18 · 168 HN pts · official
  5. ArticleQwen2.5-Max: Exploring the intelligence of large-scale MoE modelqwenlm.github.io · 2025-01-28 · official
  6. ArticleShow HN: Fine-tuned Qwen2.5-7B on 100 films for probabilistic story graphscinegraphs.ai · 2026-02-08
  7. ArticleQwen2.5-7B-Instruct-1M and Qwen2.5-14B-Instruct-1Msimonwillison.net · 2025-01-26
  8. ArticleQwen2.5-Coder Beats GPT-4o and Claude 3.5 Sonnet in Codingqwenlm.github.io · 2024-11-11 · 23 HN pts · official
  9. ArticleCivic-SLM is a domain-specialized fine-tune of Qwen2.5-7B for U.S. govt dataitsmeduncan.com · 2026-04-25
  10. ArticleQwen2.5-VLqwenlm.github.io · 2025-01-27 · official
  11. ArticleQwen2.5 Omni: See, Hear, Talk, Write, Do It Allqwenlm.github.io · 2025-03-26 · official
  12. ArticleAlibaba Qwen2.5-Omni-7B: Open Source End-to-End Multimodal AI Modelalizila.com · 2025-03-27
  13. ArticleQwen2.5-VL-32B: Smarter and Lightersimonwillison.net · 2025-03-25
  14. ArticleQwen2.5-Max: The Next Leap in AI Innovationmedium.com · 2025-01-29
  15. ArticleQwen2.5-Omni Technical Reporthuggingface.co · 2025-03-30
  16. PaperQwen2.5-Coder Technical Reportarxiv.org · 2024-11-11 · 4 HN pts
  17. DiscussionQwen2.5-coder on the events occurring during the Tiananmen Square protestsnews.ycombinator.com · 2024-11-20 · 7 HN pts
  18. CodeQwen2.5-VL: State-of-the-art multimodal LLMgithub.com · 2025-01-30
  19. DiscussionNew Qwen2.5-Max Outperforms DeepSeek V3 in Benchmarkstwitter.com · 2025-01-28
  20. ArticleGuide how to use a local LLM Qwen2.5-coder-7B as an autocompelte in VSCodeeamag.me · 2024-10-29 · 4 HN pts
  21. ArticleQwen2.5-Pluschat.qwenlm.ai · 2025-01-31
  22. ArticleShow HN: I made a free study tool powered by Qwen2.5freeflashcardgenerator.com · 2025-01-28 · 3 HN pts

Where to run it

Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.

#Cheapest live rowIn /1MOut /1MContextQuantSeen
1 Featherless — — — — Oct 25

All 1 provider →

Where it's used

Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.

Served by 1 API provider; 24 quantized builds for local use; a big base for builders (69 fine-tunes, adapters and merges).

  • As a service1Featherless
  • On own hardware24quantized builds
  • Built on6943 fine-tunes · 9 adapters · 17 merges
  • In apps0Hugging Face Spaces

Will it fit?

Weights only, an estimate: leave room for the KV cache. Method

Fits a 64 GB Mac (unified memory) at INT4

About 20 GB at INT4, 39 GB at FP8, 79 GB at BF16: weights only, so leave room for the KV cache.

16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 20 GB of weights (estimate; KV cache and activations extra)INT420 GBFP8: about 39 GB of weights (estimate; KV cache and activations extra)FP839 GBBF16: about 79 GB of weights (estimate; KV cache and activations extra)BF1679 GB
16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 20 GB of weights (estimate; KV cache and activations extra)INT420 GBFP8: about 39 GB of weights (estimate; KV cache and activations extra)FP839 GBBF16: about 79 GB of weights (estimate; KV cache and activations extra)BF1679 GB

Estimate params × bytes per param × 1.2 · how we estimate

The story so far

Weights, every dated article, and repo commits on one rail.

Most coverage came in the first two weeks

Teal is the weights (2024-11-08), amber triangles are 10 dated articles (each links to it).

OctNovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOctQwen2.5: A Party of Foundation Models · 2024-09-18Qwen2.5-Coder-32B is an LLM that can code well that runs on my Mac · 2024-11-13Show HN: IttyBittyAI – Qwen2.5:0.5B running on a 2017 Asus smartphone · 2026-09-24Qwen2.5-Coder Beats GPT-4o and Claude 3.5 Sonnet in Coding · 2024-11-11Qwen2.5-Coder-32B with coding abilities matching those of GPT-4o · 2024-11-09Qwen2.5-Coder Technical Report · 2024-11-11Qwen2.5-coder on the events occurring during the Tiananmen Square protests · 2024-11-20Guide how to use a local LLM Qwen2.5-coder-7B as an autocompelte in VSCode · 2024-10-29Qwen2.5 Coder · 2024-11-11Show HN: I made a free study tool powered by Qwen2.5 · 2025-01-28Weights 2024-11-08 · 2024-11-08Weights 2024-11-08now · 1 provider
OctNovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOctQwen2.5: A Party of Foundation Models · 2024-09-18Qwen2.5-Coder-32B is an LLM that can code well that runs on my Mac · 2024-11-13Show HN: IttyBittyAI – Qwen2.5:0.5B running on a 2017 Asus smartphone · 2026-09-24Qwen2.5-Coder Beats GPT-4o and Claude 3.5 Sonnet in Coding · 2024-11-11Qwen2.5-Coder-32B with coding abilities matching those of GPT-4o · 2024-11-09Qwen2.5-Coder Technical Report · 2024-11-11Qwen2.5-coder on the events occurring during the Tiananmen Square protests · 2024-11-20Guide how to use a local LLM Qwen2.5-coder-7B as an autocompelte in VSCode · 2024-10-29Qwen2.5 Coder · 2024-11-11Show HN: I made a free study tool powered by Qwen2.5 · 2025-01-28Weights 2024-11-08 · 2024-11-08Weights 2024-11-08now · 1 provider

What people built on it

93 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.

93 derivatives on the Hub, led by 43 fine-tunes

From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.

24 quantized (Hugging Face model tree)43 fine-tunes (Hugging Face model tree)17 merges (Hugging Face model tree)9 adapters (Hugging Face model tree)24 quantized43 fine-tunes17 merges9 adapters
24 quantized (Hugging Face model tree)43 fine-tunes (Hugging Face model tree)17 merges (Hugging Face model tree)9 adapters (Hugging Face model tree)24 quantized43 fine-tunes17 merges9 adapters
GGUF lmstudio-community/Qwen2.5-Coder-32B-GGUF for llama.cpp, Ollama, LM Studio 131.1k downloads MLX mlx-community/Qwen2.5-Coder-32B-4bit for Apple silicon 1.7k downloads GPTQ ModelCloud/Qwen2.5-Coder-32B-Instruct-gptqmodel-4bit-vortex-v1 for vLLM, SGLang 3k downloads

All variants →

Reference

Everything else, folded: the full tables, every source.

Use-case evidence

2 use cases

What it is good at 2 use cases

Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.

Facts

15 fields

Listed in LLMs

Facts

Pipeline tagtext-generation
Librarytransformers
ArchitecturesQwen2ForCausalLM
Model typeqwen2
Params by dtypeBF16: 32,763,876,352
Size bandL
MoEno
PapersarXiv:2409.12186, arXiv:2309.00071, arXiv:2407.10671
LicenseApache 2.0 (apache-2.0)
Hub snapshot2026-10-25

Hardware fit

3 estimates

Hardware fit ESTIMATE

Weights-only memory for Qwen2.5-Coder-32B: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)79 GB · Single high-end GPU (<80 GB)
FP8 (1 byte/param)39 GB · Single high-end GPU (<80 GB)
INT4 (0.5 bytes/param)20 GB · Single GPU (<24 GB)

Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.

All variants

40 variants

Lineage

Fine-tuned from Qwen/Qwen2.5-32B (same organization) Family: Qwen/Qwen2.5 93 derivatives on the Hub

Base model: Qwen/Qwen2.5-32B

quantized: 24finetune: 43merge: 17adapter: 9

Top variants (40 shown, ranked by downloads)

Relation
Quant
40 rows
Variant Relation Quant Params Band License Author Likes Downloads
lmstudio-community/Qwen2.5-Coder-32B-GGUF quantized GGUF — — apache-2.0 lmstudio-community 6 131.1k
bartowski/Qwen2.5-Coder-32B-GGUF quantized GGUF — — apache-2.0 bartowski 18 60k
ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-GGUF quantized GGUF — — apache-2.0 ThomasBaruzier 1 31.8k
unsloth/Qwen2.5-Coder-32B finetune — 32.8B L apache-2.0 unsloth 3 28.5k
unsloth/Qwen2.5-Coder-32B-bnb-4bit quantized -bnb 33.7B L apache-2.0 unsloth 6 13.7k
mradermacher/Qwen2.5-Coder-32B-i1-GGUF quantized GGUF — — apache-2.0 mradermacher 1 7.5k
ModelCloud/Qwen2.5-Coder-32B-Instruct-gptqmodel-4bit-vortex-v1 quantized gptq 32.8B L apache-2.0 ModelCloud 16 3k
mradermacher/Qwen2.5-Coder-32B-GGUF quantized GGUF — — apache-2.0 mradermacher 0 3.7k
tensorblock/Qwen2.5-Coder-32B-GGUF quantized GGUF — — apache-2.0 tensorblock 0 3.2k
YOYO-AI/Qwen2.5-32B-YOYO-V2 merge — 32.8B L apache-2.0 YOYO-AI 1 2.5k
Snowflake/Qwen-2.5-coder-Arctic-ExCoT-32B finetune — 32.8B L cc-by-nc-4.0 Snowflake 16 1.1k
mlx-community/Qwen2.5-Coder-32B-4bit quantized -4bit 32.8B L apache-2.0 mlx-community 3 1.7k
mlx-community/Qwen2.5-Coder-32B-8bit quantized -8bit 32.8B L apache-2.0 mlx-community 4 1.6k
ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2 finetune EXL2 — — apache-2.0 ThomasBaruzier 0 1.5k
nisten/deepseek-r1-qwen32b-mlx-6bit quantized mlx 32.8B L mit nisten 14 655
xxris/Qwen2.5-Coder-32B-GGUF quantized GGUF — — apache-2.0 xxris 0 1.1k
YOYO-AI/QwQ-coder-32B merge — 32.8B L — YOYO-AI 7 760
itlwas/Qwen2.5-Coder-32B-Q4_K_M-GGUF quantized GGUF — — apache-2.0 itlwas 0 985
zombiegirlcz/Qwen2.5-Coder-32B-HECK-Uncensored adapter GGUF — — apache-2.0 zombiegirlcz 0 941
Sg-at-srijan-us-kg/Qwen2.5-Coder-32B-Instruct-128k-yarn finetune — 32.8B L apache-2.0 Sg-at-srijan-us-kg 0 807
mlx-community/Qwen2.5-Coder-32B-bf16 finetune -bf16 32.8B L apache-2.0 mlx-community 1 746
ehristoforu/0109-test-32b-it merge — 32.8B L — ehristoforu 15 24
asparius/qwen-32B-sdf__414 finetune — — — — asparius 0 705
Nour-Fayed/Qwen2.5-Coder-32B-Instruct finetune — 32.8B L apache-2.0 Nour-Fayed 0 605
YOYO-AI/QwQ-Coder-instruct merge — 32.8B L apache-2.0 YOYO-AI 1 545
ClaudioItaly/Qwen2.5-Coder-32B-Q3_K_M-GGUF quantized GGUF — — apache-2.0 ClaudioItaly 1 505
spdev39/Qwen2.5-Coder-32B-Instruct finetune — 32.8B L apache-2.0 spdev39 0 547
honore-250/Umucuzi-Coder finetune — 32.8B L apache-2.0 honore-250 0 540
mbetixz/Qwen2.5-Coder-32B-Instruct finetune — 32.8B L apache-2.0 mbetixz 0 475
tomasmcm/sky-t1-coder-32b-flash merge — 32.8B L apache-2.0 tomasmcm 2 364
rexprimematrix/RiShrePro finetune — 32.8B L apache-2.0 rexprimematrix 1 396
raychan0713/Qwen2.5-Coder-32B-Instruct finetune — 32.8B L apache-2.0 raychan0713 0 431
JJcs17/Qwen2.5-Coder-32B-Instruct-128k finetune — 32.8B L apache-2.0 JJcs17 0 339
YOYO-AI/YOYO-O1-32B-V2 merge — 32.8B L — YOYO-AI 3 185
Scotto2025/Qwen2.5-Coder-32B-mlx-8Bit quantized mlx 32.8B L apache-2.0 Scotto2025 0 335
AlienKevin/swe-smith-rs-base-qwen2.5-coder-32b-instruct-teacher-glm-4.6 finetune — 32.8B L apache-2.0 AlienKevin 0 306
YOYO-AI/YOYO-O1-32B merge — 32.8B L — YOYO-AI 2 198
YOYO-AI/Qwen2.5-Coder-32B-YOYO finetune — 32.8B L — YOYO-AI 2 196
Model-SafeTensors/Qwen2.5-Coder-32B-Instruct finetune — 32.8B L apache-2.0 Model-SafeTensors 0 289
BlouseJury/Qwen2.5-Coder-32B-Instruct-EXL2-4.0bpw finetune EXL2 — — apache-2.0 BlouseJury 0 286

Repository changelog

2 changes
changed in the last 7 daysNew on the watch list: 2026-10-23Other: 2026-10-24all changes

All providers

1 provider · 1 endpoint

Where to run 1 providers · 1 endpoints

Hosted endpoints observed for Qwen2.5-Coder-32B, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting.

ProviderModel refContextQuant$/M in$/M outtok/sTTFTToolsStatusObserved
Featherless
hf
 Qwen/Qwen2.5-Coder-32B — — — — — — — live 2026-10-25

Articles and coverage

26 pieces

Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.

Official2

Qwen2.5: A Party of Foundation Models
qwenlm.github.io2024-09-18officialauthorityHN front pageabout the Qwen2.5 familyHN 168 pts · 38 comments1,744 words
Qwen2.5 introduces dense decoder-only models pretrained on up to 18 trillion tokens, available in sizes from 0.5B to 72B parameters. Most variants are licensed under Apache 2.0, excluding the 3B and 72B models, and support a 128K context window with 8K generation limits. Specialized versions for coding and mathematics offer enhanced performance on benchmarks like HumanEval and MATH.written, reviewed, cited
  • Qwen2.5-Coder-7B-Instruct outperforms many larger language models across a range of programming languages and tasks.
  • Qwen2.5-Math-72B-Instruct general performance surpasses both Qwen2-Math-72B-Instruct and GPT4-o.
Qwen2.5-Coder Beats GPT-4o and Claude 3.5 Sonnet in Coding
qwenlm.github.io2024-11-11officialauthorityabout the Qwen2.5 familyHN 23 pts · 0 comments1,210 words

Analysis6

Qwen2.5-Coder-32B is an LLM that can code well that runs on my Mac
simonwillison.net2024-11-13authorityHN front pageHN 164 pts · 65 comments635 words
Alibaba released the Qwen2.5-Coder-32B-Instruct model, an Apache 2.0 licensed open-source LLM. The 32B model runs on a 64GB MacBook Pro M2 via Ollama or MLX, fetching a 20GB quantized file. On MLX, it achieved 10.016 tokens-per-sec generation speed with 32.685 GB peak memory usage.written, reviewed, cited
  • The model requires a workaround for an SSL bug when using the standard library instead of requests or httpx.
  • Running the Ollama version was too slow because the author did not have the tool compiled to use the Mac's GPU.
Qwen2.5-Coder-32B with coding abilities matching those of GPT-4o
github.com2024-11-09authorityHN 21 pts · 6 comments1,815 words
Qwen2.5-Coder Technical Report
arxiv.org2024-11-11authorityabout the Qwen2.5 familyHN 4 pts · 0 comments
Guide how to use a local LLM Qwen2.5-coder-7B as an autocompelte in VSCode
eamag.me2024-10-29about the Qwen2.5 familyHN 4 pts · 0 comments
Qwen2.5 Coder
ollama.com2024-11-11about the Qwen2.5 familyHN 3 pts · 0 comments
Show HN: I made a free study tool powered by Qwen2.5
freeflashcardgenerator.com2025-01-28about the Qwen2.5 familyHN 3 pts · 0 comments

Community2

Show HN: IttyBittyAI – Qwen2.5:0.5B running on a 2017 Asus smartphone
news.ycombinator.com2026-09-24about the Qwen2.5 familyHN 2 pts · 0 comments494 words
Qwen2.5-coder on the events occurring during the Tiananmen Square protests
news.ycombinator.com2024-11-20about the Qwen2.5 familyHN 7 pts · 4 comments
All coverage tracked for this model (26) — append-only, nothing is ever removed
TitleSourcePublishedFirst seenSignals
Show HN: IttyBittyAI – Qwen2.5:0.5B running on a 2017 Asus smartphonenews.ycombinator.com2026-09-242026-10-14hn
kev: Jev-like model built on Qwen2.5-0.5Bgithub.com2026-09-192026-10-14hn
Civic-SLM is a domain-specialized fine-tune of Qwen2.5-7B for U.S. govt dataitsmeduncan.com2026-04-252026-10-14hn
Show HN: Fine-tuned Qwen2.5-7B on 100 films for probabilistic story graphscinegraphs.ai2026-02-082026-10-14hn
Qwen2.5-Omni Technical Reporthuggingface.co2025-03-302026-10-14hn
Alibaba Qwen2.5-Omni-7B: Open Source End-to-End Multimodal AI Modelalizila.com2025-03-272026-10-14hn
Qwen2.5 Omni: See, Hear, Talk, Write, Do It Allqwenlm.github.io official2025-03-262026-10-14hn
Qwen2.5-VL-32B: Smarter and Lightersimonwillison.net2025-03-252026-10-14hn
Qwen2.5-VL-32B: Smarter and Lighterqwenlm.github.io official2025-03-242026-10-14hn
Qwen2.5-Pluschat.qwenlm.ai2025-01-312026-10-14hn
Qwen2.5-VL: State-of-the-art multimodal LLMgithub.com2025-01-302026-10-14hn
Qwen2.5-Max: The Next Leap in AI Innovationmedium.com2025-01-292026-10-14hn
New Qwen2.5-Max Outperforms DeepSeek V3 in Benchmarkstwitter.com2025-01-282026-10-14hn
Qwen2.5-Max: Exploring the intelligence of large-scale MoE modelqwenlm.github.io official2025-01-282026-10-14hn
Show HN: I made a free study tool powered by Qwen2.5freeflashcardgenerator.com2025-01-282026-10-14hn
Qwen2.5-VLqwenlm.github.io official2025-01-272026-10-14hn
Qwen2.5-7B-Instruct-1M and Qwen2.5-14B-Instruct-1Msimonwillison.net2025-01-262026-10-14hn
Qwen2.5-1M: Deploy your own Qwen with context length up to 1M tokensqwenlm.github.io official2025-01-262026-10-14hn
Qwen2.5-coder on the events occurring during the Tiananmen Square protestsnews.ycombinator.com2024-11-202026-10-14hn
Qwen2.5-Coder-32B is an LLM that can code well that runs on my Macsimonwillison.net2024-11-132026-10-14hn
Qwen2.5-Coder Technical Reportarxiv.org2024-11-112026-10-14hn
Qwen2.5 Coderollama.com2024-11-112026-10-14hn
Qwen2.5-Coder Beats GPT-4o and Claude 3.5 Sonnet in Codingqwenlm.github.io official2024-11-112026-10-14hn
Qwen2.5-Coder-32B with coding abilities matching those of GPT-4ogithub.com2024-11-092026-10-14hn
Guide how to use a local LLM Qwen2.5-coder-7B as an autocompelte in VSCodeeamag.me2024-10-292026-10-14hn
Qwen2.5: A Party of Foundation Modelsqwenlm.github.io official2024-09-182026-10-14hn

Papers

3 papers

Papers 3

arXiv links found in the model card and its tags. "Peer-reviewed" is shown only with evidence: a published venue in OpenAlex (measured) or the authors' own journal reference on arXiv (reported).

Similar models

3 models

Similar models

  • Qwen3-32B · 32.8B
    same size band (L), same task (text-generation), more providers, permissive license
  • DeepSeek-R1-Distill-Qwen-32B · 32.8B
    same size band (L), same task (text-generation), more providers, permissive license
  • QwQ-32B · 32.8B
    same size band (L), same task (text-generation), more providers, permissive license

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-14 14:08 UTC · Last verified 2026-10-25

Follow Qwen2.5-Coder-32B

Watch this model

Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.