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 › SmolLM2 family

SmolLM2-360M

Hugging Face ↗ Follow changes

HuggingFaceTB Apache 2.0 out 2 years ago

362M parameters, text in, text out, 8K tokens of context.

Verified 2026-10-25 Rank #673 in LLMs · score 72.7 method v1 Sources 3 Downloads 30d 111.4k · #310 of 619 tracked

Can I run it?0.2GB at INT4Estimatefits a 16 GB cardmethod
Can I use it?Apache 2.0Permissive; patent grantlicence file2026-10-24
Do people use it?111.4kdownloads, 30 days#310 of 619 we track · top 50%+0 likes, downloads −68% this week143 likesHugging Face2026-10-24

Worth reading

Picked by rule from 8 tracked pieces: the lab's own post, the most shared, a field report, the latest.

Most discussed · 110 HN pts

SmolLM2

SmolLM2 is a new family of compact language models from Hugging Face researchers available in three sizes ranging from 135M to 1.7B parameters.

simonwillison.net 2024-11-02 110 HN pts · 47 comments 253 words about the SmolLM2 family family
Latest · 21 months ago

Using pip to install a Large Language Model that's under 100MB

Simon Willison released llm-smollm2, a pip-installable plugin bundling the SmolLM2-135M-Instruct.Q4_1 model.

  • The author used Gemini 2.0 Pro to analyze the llama-cpp-python codebase and find logging suppression methods.
  • The 94MB version is explicitly described as not being useful for any serious tasks.
simonwillison.net 2025-02-07 24 HN pts · 5 comments 1.2k words about the SmolLM2 family family

8 best of 8 tracked

The 2 picks above, plus these 6. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

  1. CodeThe Bunny B1: Demoing a Natrual Language to App Call Interface with SmolLM2github.com · 2024-11-05 · 14 HN pts
  2. ArticleSmolLM2 (LLM) running in browser and WebGPUsimonwillison.net · 2024-11-29 · 2 HN pts
  3. PaperSmolLM2: When Smol Goes Big – Data-Centric Training of a Small Language Modelarxiv.org · 2025-02-06 · 1 HN pts
  4. DiscussionSmolLM2: The new best smol models for on-device applications (all in the open)twitter.com · 2024-11-01 · 2 HN pts
  5. ArticleHugging Face and Pruna AI: SmolLM2, Now 7x Smaller and 2x Fasterpruna.ai · 2025-01-06 · 1 HN pts
  6. Articlesmollm2ollama.com · 2024-11-01 · 1 HN pts

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.

No API provider serves it, so its downloads are people running it themselves: 43 quantized builds and 91 Spaces.

  • As a service0no API provider listed
  • On own hardware43quantized builds
  • Built on150130 fine-tunes · 19 adapters · 1 merges
  • In apps91Hugging Face Spaces

Top Spaces running it: pliny-the-prompter/obliteratus ♥ 621 · FlameF0X/SLM-Playground ♥ 4 · Kodacoda/obliteratus ♥ 2 · dumbordumber/obliteratus ♥ 2

Will it fit?

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

Fits a 16 GB card at INT4

About 0.2 GB at INT4, 0.4 GB at FP8, 0.9 GB at BF16: weights only, so leave room for the KV cache.

16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 0.2 GB of weights (estimate; KV cache and activations extra)INT40.2 GBFP8: about 0.4 GB of weights (estimate; KV cache and activations extra)FP80.4 GBBF16: about 0.9 GB of weights (estimate; KV cache and activations extra)BF160.9 GB
16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 0.2 GB of weights (estimate; KV cache and activations extra)INT40.2 GBFP8: about 0.4 GB of weights (estimate; KV cache and activations extra)FP80.4 GBBF16: about 0.9 GB of weights (estimate; KV cache and activations extra)BF160.9 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 (2024-10-31), amber triangles are 8 dated articles (each links to it), grey ticks are repo commits.

NovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOct2025-02-06: 1 repo commit2024-11-24: 1 repo commit2024-11-18: 1 repo commit2024-11-05: 1 repo commit2024-10-31: 6 repo commitsSmolLM2 · 2024-11-02Using pip to install a Large Language Model that's under 100MB · 2025-02-07The Bunny B1: Demoing a Natrual Language to App Call Interface with SmolLM2 · 2024-11-05SmolLM2 (LLM) running in browser and WebGPU · 2024-11-29SmolLM2: When Smol Goes Big – Data-Centric Training of a Small Language Model · 2025-02-06SmolLM2: The new best smol models for on-device applications (all in the open) · 2024-11-01smollm2 · 2024-11-01Hugging Face and Pruna AI: SmolLM2, Now 7x Smaller and 2x Faster · 2025-01-06Weights 2024-10-31 · 2024-10-31Weights 2024-10-31now
NovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOct2025-02-06: 1 repo commit2024-11-24: 1 repo commit2024-11-18: 1 repo commit2024-11-05: 1 repo commit2024-10-31: 6 repo commitsSmolLM2 · 2024-11-02Using pip to install a Large Language Model that's under 100MB · 2025-02-07The Bunny B1: Demoing a Natrual Language to App Call Interface with SmolLM2 · 2024-11-05SmolLM2 (LLM) running in browser and WebGPU · 2024-11-29SmolLM2: When Smol Goes Big – Data-Centric Training of a Small Language Model · 2025-02-06SmolLM2: The new best smol models for on-device applications (all in the open) · 2024-11-01smollm2 · 2024-11-01Hugging Face and Pruna AI: SmolLM2, Now 7x Smaller and 2x Faster · 2025-01-06Weights 2024-10-31 · 2024-10-31Weights 2024-10-31now

What people built on it

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

193 derivatives on the Hub, most of them fine-tunes

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

43 quantized (Hugging Face model tree)130 fine-tunes (Hugging Face model tree)1 merge (Hugging Face model tree)19 adapters (Hugging Face model tree)43 quantized130 fine-tunes1 merge19 adapters
43 quantized (Hugging Face model tree)130 fine-tunes (Hugging Face model tree)1 merge (Hugging Face model tree)19 adapters (Hugging Face model tree)43 quantized130 fine-tunes1 merge19 adapters
GGUF mradermacher/SmolLM2-360M-i1-GGUF for llama.cpp, Ollama, LM Studio 5.2k downloads

All variants →

Reference

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

Benchmarks

17 rows

Benchmarks 17 rows

measured third-party leaderboard rowauthor-reported number from the model card or eval files (17)No confidence intervals unless the source gave them; settings differ between rows, so compare within a benchmark, not across. Date = the benchmark's own date when it gives one, else when we read it.
BenchmarkValueSetting / notesReported byDate
author-reported ARC (Average)53Base Pre-Trained Modelmodel author2026-10-07
author-reported ARC (Average)43.7Instruction Modelmodel author2026-10-07
author-reported BBH (3-shot)27.3Instruction Modelmodel author2026-10-07
author-reported Commonsense QA sets · Chat and follow instructions71.7Base Pre-Trained Modelmodel author2026-10-07
author-reported Commonsense QA sets · Chat and follow instructions38Base Pre-Trained Modelmodel author2026-10-07
author-reported Commonsense QA sets · Chat and follow instructions16.9Base Pre-Trained Modelmodel author2026-10-07
author-reported Commonsense QA sets · Chat and follow instructions37.4Base Pre-Trained Modelmodel author2026-10-07
author-reported Commonsense QA sets · Chat and follow instructions70.8Instruction Modelmodel author2026-10-07
author-reported GSM8K · Reason through hard problems3.2Base Pre-Trained Modelmodel author2026-10-07
author-reported GSM8K · Reason through hard problems7.43Instruction Modelmodel author2026-10-07
author-reported HellaSwag · Chat and follow instructions54.5Base Pre-Trained Modelmodel author2026-10-07
author-reported HellaSwag · Chat and follow instructions52.1Instruction Modelmodel author2026-10-07
author-reported IFEval · Chat and follow instructions41Instruction Modelmodel author2026-10-07
author-reported MMLU · Reason through hard problems35.8Base Pre-Trained Modelmodel author2026-10-07
author-reported MMLU · Reason through hard problems32.8Instruction Modelmodel author2026-10-07
author-reported MT-Bench · Chat and follow instructions3.66Instruction Modelmodel author2026-10-07
author-reported Winogrande · Chat and follow instructions52.5Base Pre-Trained Modelmodel author2026-10-07

What the card says

1 cited sentence

Summary generated, cited

The model uses a transformer decoder architecture and is released under the Apache 2.0 license.[1][2]

  1. “Architecture: Transformer decoder”
  2. license

Generated 2026-10-07 15:11 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.

Use-case evidence

4 use cases

What it is good at 4 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

21 fields

Listed in LLMs

Facts

Pipeline tagtext-generation
Librarytransformers
ArchitecturesLlamaForCausalLM
Model typellama
Dtypebfloat16
Params by dtypeBF16: 361,821,120
Size bandS
MoEno
Context length8,192 (config.json max_position_embeddings)
Languagesen
PapersarXiv:2502.02737
LicenseApache 2.0 (apache-2.0)
Hub snapshot2026-10-25

Servability

vLLM · SGLang

Servability

vLLM: supported · LlamaForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · LlamaForCausalLM · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed

vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.

Hardware fit

3 estimates

Hardware fit ESTIMATE

Weights-only memory for SmolLM2-360M: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)0.9 GB · Edge (<4 GB)
FP8 (1 byte/param)0.4 GB · Edge (<4 GB)
INT4 (0.5 bytes/param)0.2 GB · Edge (<4 GB)

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

All variants

40 variants

Lineage

Original model Family: HuggingFaceTB/SmolLM2 193 derivatives on the Hub

quantized: 43finetune: 130merge: 1adapter: 19

Top variants (40 shown, ranked by downloads)

Relation
Size
40 rows
Variant Relation Quant Params Band License Author Likes Downloads
HuggingFaceTB/SmolLM2-360M-Instruct quantized — 362M S apache-2.0 HuggingFaceTB 235 8M
saheedniyi/YarnGPT2 finetune — 366M S apache-2.0 saheedniyi 12 107.7k
unsloth/SmolLM2-360M finetune — 362M S apache-2.0 unsloth 1 40k
saheedniyi/YarnGPT finetune — 366M S apache-2.0 saheedniyi 54 8.3k
mradermacher/SmolLM2-360M-i1-GGUF quantized GGUF — — apache-2.0 mradermacher 0 5.2k
prithivMLmods/SmolLM2-360M-GGUF quantized GGUF — — creativeml-openrail-m prithivMLmods 1 4.7k
ReDiX/SmolLM2-360M-Instruct-ita finetune — — — apache-2.0 ReDiX 0 4.3k
mags0ft/SmolLM2-360m-German-Instruct quantized GGUF 362M S apache-2.0 mags0ft 2 3.2k
assskelad/smollm2-360M-sft finetune — 362M S — assskelad 0 3.2k
saheedniyi/YarnGPT2b finetune — 366M S apache-2.0 saheedniyi 5 2.9k
qgallouedec/SmolLM2-360M-Rickified finetune — 362M S — qgallouedec 0 3k
saheedniyi/YarnGPT-local finetune — 366M S cc-by-nc-sa-4.0 saheedniyi 18 1.9k
mradermacher/SmolLM2-360M-GGUF quantized GGUF — — apache-2.0 mradermacher 0 2.8k
tensorblock/SmolLM2-360M-GGUF quantized GGUF — — apache-2.0 tensorblock 1 2.6k
davanstrien/Smol-Hub-tldr finetune — 362M S mit davanstrien 11 2k
Mungert/SmolLM2-360M-Instruct-GGUF quantized GGUF — — apache-2.0 Mungert 0 2.4k
levzalt/canter finetune — 2.06B M other levzalt 0 2k
onnx-community/SmolLM2-360M-ONNX quantized — — — apache-2.0 onnx-community 1 1.8k
yava-code/Tessera-1B-Nano finetune — 383M S apache-2.0 yava-code 1 1.6k
yava-code/Tessera-1B-Nano-Base finetune — 362M S apache-2.0 yava-code 1 1.5k
axie66/smollm2-360M-sft finetune — 362M S — axie66 0 1.2k
Aravindan/smol-lm2-360-instruct finetune — 362M S apache-2.0 Aravindan 3 918
argilla/SmolLM2-360M-synthetic-concise-reasoning finetune — 362M S apache-2.0 argilla 2 723
vonjack/SmolLM2-360M-Merged merge GGUF 362M S — vonjack 0 814
hzxie/dynamic-vla-DOM finetune — 430M S other hzxie 0 754
lewtun/SmolLM2-360M-SFT-everyday-conversations finetune — 362M S — lewtun 0 655
isotnek/SmolLM2-360M-Instruct-heretic finetune — 362M S apache-2.0 isotnek 0 584
itlwas/SmolLM2-360M-Q4_K_M-GGUF quantized GGUF — — apache-2.0 itlwas 0 555
deepakshirkem/image-description_to_emotion finetune — 362M S apache-2.0 deepakshirkem 1 492
fausap/peft-smollm2-lora-gtx1660 adapter — — — apache-2.0 fausap 0 531
SaeedAngiz1/shahnameh-smollm2-360m finetune — 362M S apache-2.0 SaeedAngiz1 1 470
punkt2/foodcalculator-nutrition-SmolLM2-360M finetune — 362M S — punkt2 0 478
nimendraai/SmolLM2-360M-Assignment-Metadata-Extractor quantized GGUF — — apache-2.0 nimendraai 1 427
Michaelj1/finetune-smolLM2-360M finetune — 362M S apache-2.0 Michaelj1 0 463
CommandCode/agr-flash finetune — — — apache-2.0 CommandCode 4 194
dianak12/SmolLM2-360M-emotions-oneshot quantized GGUF — — apache-2.0 dianak12 0 388
gustavokuklinski/aeon-360M finetune — 362M S mit gustavokuklinski 1 297
Edgeai-gio/SmolCpp-MoE-360M quantized GGUF — — apache-2.0 Edgeai-gio 2 243
voidful/smol-360m-ft finetune — 362M S apache-2.0 voidful 0 341
assskelad/smollm2-360M-sft_SmallThoughts finetune — 362M S — assskelad 0 326

Repository changelog

10 commits

Repository changelog 10 commits

DateCommitSHA
2025-02-06 Update README.md f8027fd
2024-11-24 Update README.md 426f83c
2024-11-18 Update README.md d9ced15
2024-11-05 Update config.json 3ce05f6
2024-10-31 Update README.md f765b5d
2024-10-31 Update README.md 6f6fbd4
2024-10-31 Update README.md 61c4543
2024-10-31 Upload tokenizer f98257b
2024-10-31 Upload LlamaForCausalLM ef821dc
2024-10-31 initial commit 751d158

Articles and coverage

8 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.

Analysis7

SmolLM2
simonwillison.net2024-11-02authorityHN front pageabout the SmolLM2 familyHN 110 pts · 47 comments253 words
SmolLM2 is a new family of compact language models from Hugging Face researchers available in three sizes ranging from 135M to 1.7B parameters. Trained on 11 trillion tokens using FineWeb-Edu, DCLM, and The Stack, these lightweight models are designed to run on-device and solve a wide range of tasks efficiently. The model weights are released under an Apache 2 license.written, reviewed, cited
Using pip to install a Large Language Model that's under 100MB
simonwillison.net2025-02-07authorityabout the SmolLM2 familyHN 24 pts · 5 comments1,165 words
Simon Willison released llm-smollm2, a pip-installable plugin bundling the SmolLM2-135M-Instruct.Q4_1 model. The package is 94MB to comply with PyPI limits, utilizing the llama-cpp-python library with verbose=False and logging.CRITICAL to suppress output. While the model is not useful for serious tasks, it serves as a functional example of running a quantized LLM via pip.written, reviewed, cited
  • The author used Gemini 2.0 Pro to analyze the llama-cpp-python codebase and find logging suppression methods.
  • The 94MB version is explicitly described as not being useful for any serious tasks.
The Bunny B1: Demoing a Natrual Language to App Call Interface with SmolLM2
github.com2024-11-05authorityabout the SmolLM2 familyHN 14 pts · 0 comments265 words
The Bunny B1 demo combines SmolLM2 with Outlines to create a natural language interface for local apps on smaller devices. It requires a Rust compiler and allows users to edit functions.json to add new actions like ordering food or rides. The model defaults to the mps device on Macs but can be swapped for cuda.written, reviewed, cited
SmolLM2 (LLM) running in browser and WebGPU
simonwillison.net2024-11-29authorityabout the SmolLM2 familyHN 2 pts · 0 comments222 words
Hugging Face's SmolLM2-1.7B-Instruct runs directly in a web browser using WebGPU to perform structured text extraction. The demo relies on web-llm by MLC and its XGrammar library for JSON schema and EBNF-based guidance, requiring Chrome for execution. It loads 924.8MB of model data to convert image descriptions into GitHub issues defined by JSON schema.written, reviewed, cited
SmolLM2: When Smol Goes Big – Data-Centric Training of a Small Language Model
arxiv.org2025-02-06authorityabout the SmolLM2 familyHN 1 pts · 0 comments368 words
SmolLM2 is a 1.7 billion parameter language model that outperforms recent small LMs including Qwen2.5-1.5B and Llama3.2-1B. The model trains on ~11 trillion tokens using a multi-stage process that mixes web text with specialized math, code, and instruction-following data. New specialized datasets named FineMath, Stack-Edu, and SmolTalk were introduced to address limitations in existing data quality and size.written, reviewed, cited
smollm2
ollama.com2024-11-01about the SmolLM2 familyHN 1 pts · 0 comments73 words
Hugging Face and Pruna AI: SmolLM2, Now 7x Smaller and 2x Faster
pruna.ai2025-01-06about the SmolLM2 familyHN 1 pts · 0 comments192 words

Community1

SmolLM2: The new best smol models for on-device applications (all in the open)
twitter.com2024-11-01about the SmolLM2 familyHN 2 pts · 0 comments

Papers

1 paper

Papers 1

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

  • LFM2.5-350M · 354M
    same size band (S), same task (text-generation)
  • gpt2-medium · 380M
    same size band (S), same task (text-generation), permissive license
  • Qwen2.5-0.5B-Instruct · 494M
    same size band (S), same task (text-generation), more providers, permissive license

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:40 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:11 UTC · Last verified 2026-10-25

Follow SmolLM2-360M

Watch this model

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