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 › Apertus family

Apertus-8B-Instruct-2509

Hugging Face ↗ Follow changes

swiss-ai Apache 2.0 out 14 months ago

8.05B parameters, text in, text out, 64K tokens of context. 2 providers serve it from $0.10 per million input tokens.

Verified 2026-10-25 Rank #221 in LLMs · score 87.8 method v1 Sources 3 Downloads 30d 420.7k · #179 of 619 tracked

Changed this week: 1 new provider: Featherless.2026-10-23

Can I run it?4.8GB at INT4Estimatefits a 16 GB cardmethod
What does it cost?$0.10/1M incheapest of 1Public AI2026-10-25
Can I use it?Apache 2.0Permissive; patent grantlicence file2026-10-24
Do people use it?420.7kdownloads, 30 days#179 of 619 we track · top 29%+3 likes, downloads −4% this week499 likesHugging Face2026-10-24

Worth reading

Picked by rule from 23 tracked pieces: the lab's own post, the most shared, a field report, the latest.

23 best of 23 tracked

The 3 picks above, plus these 20. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

  1. ArticleApertus 1.5, Swiss open-weight, open-source, open training data modelapertus-ai.org · 2026-07-25 · 10 HN pts
  2. ArticleApertus 1.5 out – Latest version of Switzerland's open model with 70B versioncscs.ch · 2026-07-24 · 7 HN pts
  3. ArticleSwiss public broadcaster to block AI scrapers – except Swiss model Apertusnzz.ch · 2026-08-12 · 3 HN pts
  4. ArticleApertus – Open-Source Cameraapertus.org · 2023-11-26 · 184 HN pts
  5. PaperApertus: Democratizing Open and Compliant LLMs for ...arxiv.org · undated
  6. ArticleApertus 8B and 70B – a new open multilingual LLM from Switzerlandactu.epfl.ch · 2025-09-02 · 71 HN pts
  7. ArticleApertus: An open, transparent, multilingual language modelethz.ch · 2025-09-02 · 16 HN pts
  8. PaperApertus: Democratizing Open and Compliant LLMs for Global Language Environmentsarxiv.org · 2025-09-22
  9. ArticleApertus: A fully open, transparent, multilingual language modelswisscom.ch · 2025-09-02
  10. ArticleApertus includes many languages that have so far been underrepresented in LLMsswiss-ai.org · 2025-09-02
  11. ArticlePublic Summary of Training Content for General-Purpose AI modelsaial.ie · undated
  12. ArticleApertus-8B-Instruct-2509 Benchmark Results · The Aggregatetheaggregate.ai · undated
  13. Article[Literature Review] Apertus: Democratizing Open and Compliant LLMs for Global Language Environmentsthemoonlight.io · undated
  14. ArticleApertus, the most powerful open-source LLM released by a public institutionpublicai.co · 2025-09-05
  15. ArticleApertus: Open Source Cinemaapertus.org · 2017-09-24
  16. ArticleApertus, an Open Source Cinema Camerablog.makezine.com · 2012-02-12
  17. DiscussionApertus - open source cinema Newsnews.ycombinator.com · 2011-10-29
  18. ArticleApertus: An open, transparent, multilingual language modelrobohub.org · 2025-09-16
  19. ArticleLLM Benchmark Evaluation - Apertus-8B | DS-NLP Labblog.nlp-lab.ai · 2025-09-05
  20. ArticleApertus 8B Instruct benchmark scores — Benchmark Atlasatlas.kevinhu.io · 2025-09-02

Where to run it

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

#Cheapest 2In /1MOut /1MContextQuantSeen
1 Public AI $0.10 $0.20 — — Oct 25
2 Featherless — — — — Oct 25

All 2 providers →

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 2 API providers; 39 quantized builds for local use; a big base for builders (146 fine-tunes, adapters and merges).

  • As a service2Featherless, Public AI
  • On own hardware39quantized builds
  • Built on14620 fine-tunes · 124 adapters · 2 merges
  • In apps26Hugging Face Spaces

Top Spaces running it: akhaliq/Apertus-8B-Instruct-2509 ♥ 17 · loleg/fastapi-apertus ♥ 1 · Rudran/apertus ♥ 1

Will it fit?

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

Fits a 16 GB card at INT4

About 4.8 GB at INT4, 9.7 GB at FP8, 19 GB at BF16: weights only, so leave room for the KV cache.

16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 4.8 GB of weights (estimate; KV cache and activations extra)INT44.8 GBFP8: about 9.7 GB of weights (estimate; KV cache and activations extra)FP89.7 GBBF16: about 19 GB of weights (estimate; KV cache and activations extra)BF1619 GB
16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 4.8 GB of weights (estimate; KV cache and activations extra)INT44.8 GBFP8: about 9.7 GB of weights (estimate; KV cache and activations extra)FP89.7 GBBF16: about 19 GB of weights (estimate; KV cache and activations extra)BF1619 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.

8 dated articles and 20 repo commits so far

Teal is the weights (2025-08-13), amber triangles are 8 dated articles (each links to it), grey ticks are repo commits.

DecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOct2026-07-17: 1 repo commit2026-06-20: 1 repo commit2026-04-21: 1 repo commit2025-11-14: 1 repo commit2025-10-01: 1 repo commit2025-09-30: 1 repo commit2025-09-05: 1 repo commit2025-09-03: 2 repo commits2025-09-02: 6 repo commits2025-09-01: 2 repo commits2025-08-30: 1 repo commit2025-08-13: 2 repo commitsApertus – Open Foundation Model for Sovereign AI · 2026-06-21Apertus 1.5, Swiss open-weight, open-source, open training data model · 2026-07-25Proton partners with Apertus, Switzerland's sovereign AI model · 2026-09-18Apertus 1.5 out – Latest version of Switzerland's open model with 70B version · 2026-07-24Swiss public broadcaster to block AI scrapers – except Swiss model Apertus · 2026-08-12Apertus – Open-Source Camera · 2023-11-26Apertus 8B and 70B – a new open multilingual LLM from Switzerland · 2025-09-02Apertus: An open, transparent, multilingual language model · 2025-09-02Weights 2025-08-13 · 2025-08-13Weights 2025-08-13now · 2 providers
DecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOct2026-07-17: 1 repo commit2026-06-20: 1 repo commit2026-04-21: 1 repo commit2025-11-14: 1 repo commit2025-10-01: 1 repo commit2025-09-30: 1 repo commit2025-09-05: 1 repo commit2025-09-03: 2 repo commits2025-09-02: 6 repo commits2025-09-01: 2 repo commits2025-08-30: 1 repo commit2025-08-13: 2 repo commitsApertus – Open Foundation Model for Sovereign AI · 2026-06-21Apertus 1.5, Swiss open-weight, open-source, open training data model · 2026-07-25Proton partners with Apertus, Switzerland's sovereign AI model · 2026-09-18Apertus 1.5 out – Latest version of Switzerland's open model with 70B version · 2026-07-24Swiss public broadcaster to block AI scrapers – except Swiss model Apertus · 2026-08-12Apertus – Open-Source Camera · 2023-11-26Apertus 8B and 70B – a new open multilingual LLM from Switzerland · 2025-09-02Apertus: An open, transparent, multilingual language model · 2025-09-02Weights 2025-08-13 · 2025-08-13Weights 2025-08-13now · 2 providers

What people built on it

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

185 derivatives on the Hub, most of them adapters

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

39 quantized (Hugging Face model tree)20 fine-tunes (Hugging Face model tree)2 merges (Hugging Face model tree)124 adapters (Hugging Face model tree)39 quantized20 fine-tunes2 merges124 adapters
39 quantized (Hugging Face model tree)20 fine-tunes (Hugging Face model tree)2 merges (Hugging Face model tree)124 adapters (Hugging Face model tree)39 quantized20 fine-tunes2 merges124 adapters
GGUF unsloth/Apertus-8B-Instruct-2509-GGUF for llama.cpp, Ollama, LM Studio 81.1k downloads MLX mlx-community/Apertus-8B-Instruct-2509-4bit for Apple silicon 3.5k downloads FP8 RedHatAI/Apertus-8B-Instruct-2509-FP8-dynamic for vLLM, SGLang 164.1k downloads

All variants →

Reference

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

Benchmarks

15 rows

Benchmarks 15 rows

measured third-party leaderboard rowauthor-reported number from the model card or eval files (15)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 Challenge · Chat and follow instructions72.7 percentApertus-8Bmodel author2026-10-07
author-reported ARC Challenge · Chat and follow instructions70.6 percentApertus-70Bmodel author2026-10-07
author-reported Avg65.8 percentApertus-8Bmodel author2026-10-07
author-reported Avg67.5 percentApertus-70Bmodel author2026-10-07
author-reported Commonsense QA sets · Chat and follow instructions79.8 percentApertus-8Bmodel author2026-10-07
author-reported Commonsense QA sets · Chat and follow instructions81.9 percentApertus-70Bmodel author2026-10-07
author-reported HellaSwag · Chat and follow instructions59.8 percentApertus-8Bmodel author2026-10-07
author-reported HellaSwag · Chat and follow instructions64 percentApertus-70Bmodel author2026-10-07
author-reported Winogrande · Chat and follow instructions70.6 percentApertus-8Bmodel author2026-10-07
author-reported Winogrande · Chat and follow instructions73.3 percentApertus-70Bmodel author2026-10-07
author-reported XCOPA · Work across languages66.5 percentApertus-8Bmodel author2026-10-07
author-reported XCOPA · Work across languages69.8 percentApertus-70Bmodel author2026-10-07
author-reported XNLI · Classify and make typed decisions45.2 percentApertus-8Bmodel author2026-10-07
author-reported XNLI · Classify and make typed decisions45.3 percentApertus-70Bmodel author2026-10-07
author-reported open_question (LEXam-Benchmark/LEXam)22.44—LEXam Leaderboard2026-10-24

Use-case evidence

6 use cases

What it is good at 6 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
ArchitecturesApertusForCausalLM
Model typeapertus
Dtypebfloat16
Params by dtypeBF16: 8,053,338,176
Size bandM
MoEno
Context length65,536 (config.json max_position_embeddings)
PapersarXiv:2509.14233
LicenseApache 2.0 (apache-2.0)
Hub snapshot2026-10-25

Servability

vLLM · SGLang

Servability

vLLM: supported · ApertusForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · ApertusForCausalLM · 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 Apertus-8B-Instruct-2509: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)19 GB · Single GPU (<24 GB)
FP8 (1 byte/param)9.7 GB · Single GPU (<24 GB)
INT4 (0.5 bytes/param)4.8 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 swiss-ai/Apertus-8B-2509 (same organization) Family: swiss-ai/Apertus 185 derivatives on the Hub

Base model: swiss-ai/Apertus-8B-2509

quantized: 39finetune: 20merge: 2adapter: 124

Top variants (40 shown, ranked by downloads)

Relation
Quant
Size
40 rows
Variant Relation Quant Params Band License Author Likes Downloads
RedHatAI/Apertus-8B-Instruct-2509-FP8-dynamic quantized -FP8 8.05B M apache-2.0 RedHatAI 6 164.1k
unsloth/Apertus-8B-Instruct-2509-GGUF quantized GGUF — — apache-2.0 unsloth 23 81.1k
aisingapore/Apertus-SEA-LION-v4-8B-IT-GGUF finetune GGUF — — apache-2.0 aisingapore 1 37.7k
bartowski/swiss-ai_Apertus-8B-Instruct-2509-GGUF quantized GGUF — — — bartowski 4 15.9k
aisingapore/Apertus-SEA-LION-v4-8B-IT finetune — 8.05B M mit aisingapore 5 9k
EPFLiGHT/Apertus-8B-MeditronFO finetune — 8.05B M apache-2.0 EPFLiGHT 9 8.3k
DevQuasar/swiss-ai.Apertus-8B-Instruct-2509-GGUF quantized GGUF — — — DevQuasar 3 8.4k
mradermacher/Apertus-8B-Instruct-2509-i1-GGUF quantized GGUF — — apache-2.0 mradermacher 0 8.5k
mlx-community/Apertus-8B-Instruct-2509-bf16 finetune -bf16 8.05B M apache-2.0 mlx-community 6 5.4k
giladgd/Apertus-8B-Instruct-2509-GGUF quantized GGUF — — apache-2.0 giladgd 0 5.4k
mradermacher/Apertus-8B-Instruct-2509-GGUF quantized GGUF — — apache-2.0 mradermacher 0 5.2k
mlx-community/Apertus-8B-Instruct-2509-4bit quantized -4bit 8.05B M apache-2.0 mlx-community 1 3.5k
redponike/Apertus-8B-Instruct-2509-GGUF quantized GGUF — — apache-2.0 redponike 1 2.6k
tartuNLP/Apertus-EstLLM-8B-Instruct-0326 merge — 8.05B M apache-2.0 tartuNLP 1 2.1k
sevri/Apertus-8B-Instruct-2509-W8A8-CALIBRATED quantized -W8A8 8.05B M — sevri 0 2.2k
anicka/karma-electric-apertus-8b quantized GGUF 8.05B M apache-2.0 anicka 0 2.1k
mlx-community/Apertus-8B-Instruct-2509-8bit quantized -8bit 8.05B M apache-2.0 mlx-community 1 1.9k
MaziyarPanahi/Apertus-8B-Instruct-2509-GGUF quantized GGUF — — — MaziyarPanahi 0 1.6k
jgerster0/Apertus-8B-Instruct-2509-LOGEQ-FP8_dynamic quantized — 8.05B M — jgerster0 0 1.1k
TevunahAi/Apertus-8B-Instruct-TevunahAi-GPTQ quantized GPTQ 10.8B L apache-2.0 TevunahAi 0 1.1k
mlx-community/Apertus-8B-Instruct-2509-6bit quantized — 8.05B M apache-2.0 mlx-community 1 1.1k
jgerster0/Apertus-8B-Instruct-2509-W8A16-CALIBRATED quantized — 2.82B M — jgerster0 0 1.1k
sevri/Apertus-8B-Instruct-2509-SPINQUANT-FP8_dynamic quantized — 8.05B M — sevri 0 1.1k
Colby/apertus-8b-coding-gguf quantized gguf — — — Colby 1 819
thomaskiefer/EAGLE3-Apertus-8B-Instruct-2509 finetune — 513M S apache-2.0 thomaskiefer 0 740
saidutta69/Apertus-8B-Instruct-heretic finetune GGUF 8.05B M apache-2.0 saidutta69 0 683
loleg/Apertus-8B-Instruct-2509-mlx finetune mlx 8.05B M apache-2.0 loleg 2 536
kalle07/apertus-8b_instruct-2509-heretic_R7_K005_q8_0-gguf quantized gguf — — apache-2.0 kalle07 1 535
llmat/Apertus-8B-Instruct-2509-NVFP4 quantized NVFP4 5.00B M apache-2.0 llmat 1 531
bg-digitalservices/Apertus-8B-Instruct-2509-NVFP4 quantized NVFP4 5.00B M apache-2.0 bg-digitalservices 1 507
gaparecido/apertus-8b-financial-reasoner-v1 finetune — 8.05B M apache-2.0 gaparecido 1 454
anicka/ke-v9-8b quantized GGUF 8.05B M apache-2.0 anicka 0 476
safouaneelg/Apertus-8B-Instruct-2509-Math-Step-DPO-10K quantized — 8.05B M apache-2.0 safouaneelg 0 473
bg-digitalservices/Apertus-8B-Instruct-2509-NVFP4A16 quantized — 5.00B M apache-2.0 bg-digitalservices 0 395
gaparecido/apertus-8b-financial-reasoner-v2 finetune — — — apache-2.0 gaparecido 4 177
moncefem/legitus-instruct-v1 merge — 8.05B M apache-2.0 moncefem 0 322
jondale/Apertus-8B-Instruct-2509-GGUF quantized GGUF — — apache-2.0 jondale 0 288
onnx-community/Apertus-8B-Instruct-2509-ONNX quantized — — — apache-2.0 onnx-community 1 198
austinjp/Apertus-8B-Instruct-2509-Q4_K_M-GGUF quantized GGUF — — apache-2.0 austinjp 0 191
safouaneelg/Apertus-8B-Instruct-2509-GSM8k-SFT quantized — 8.05B M apache-2.0 safouaneelg 0 188

Repository changelog

20 commits
changed in the last 7 daysNew providers: 2026-10-23all changes

Repository changelog 20 commits

DateCommitSHA
2026-07-17 Revert token 73 to <SPECIAL_73>: <|image|> belongs to the multimodal tokenizers, not this text tokenizer. The chat template keeps referencing it by name and binds against whichever tokenizer ships with an image-capable model. b946d40
2026-06-20 Add LEXam evaluation results (#32) 2256fb4
2026-04-21 Correcting the link for the report. The previous one is obsolete. (#27) 50761a5
2025-11-14 Add support for images and fix end of assistant issue (#22) cdb3e4f
2025-10-01 remove gating, add policy 5df61a7
2025-09-30 Update README.md 26fc8ea
2025-09-05 Update README.md a76c9e6
2025-09-03 Update README.md 7846a2c
2025-09-03 Upload tokenizer.json 547207e
2025-09-02 Update README.md 9579a2d
2025-09-02 Update README.md 87eb936
2025-09-02 Update README.md 6063fa1
2025-09-02 Create license b31bb85
2025-09-02 Changed model name to Apertus-8B-Instruct-2509 (#1) 36c8187
2025-09-02 initial model card abbfaae
2025-09-01 Update chat_template.jinja ffb69ff
2025-09-01 Add files using upload-large-folder tool 55ec663
2025-08-30 upload chat_template, generation and tokenizer configs fa7d504
2025-08-13 Update README.md 00e43f3
2025-08-13 initial commit bae112d

All providers

2 providers · 2 endpoints

Where to run 2 providers · 2 endpoints

Hosted endpoints observed for Apertus-8B-Instruct-2509, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting. Cheapest observed input price: $0.10/M.

ProviderModel refContextQuant$/M in$/M outtok/sTTFTToolsStatusObserved
Public AI
hf
 swiss-ai/apertus-8b-instruct — — $0.10 $0.20 93.4 4683 ms no live 2026-10-25
Featherless
hf
 swiss-ai/Apertus-8B-Instruct-2509 — — — — — — — live 2026-10-25

Articles and coverage

23 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.

Analysis10

Apertus – Open Foundation Model for Sovereign AI
apertvs.ai2026-06-21HN front pageabout the Apertus familyHN 534 pts · 183 comments94 words
Apertus 1.5, Swiss open-weight, open-source, open training data model
apertus-ai.org2026-07-25about the Apertus familyHN 10 pts · 3 comments2,079 words
Apertus 1.5 extends the 8B and 70B models with native multimodal understanding, an optional thinking mode, and a four times longer context window. Released to Hugging Face on July 24, 2026 under an Apache 2.0 license, the models feature a 262,144 token context window and use the Emu3.5 tokenizer for images and WavTokenizer for audio. The update adds 4 trillion tokens of training data to the 8B model and 2 trillion tokens to the 70B model.written, reviewed, cited
  • The architecture remains unchanged because the models only extend the vocabulary to accept discrete image and audio tokens.
  • Reasoning traces added during SFT survive subsequent text-only reinforcement learning, improving mathematics more than starting from a text-only checkpoint.
Proton partners with Apertus, Switzerland's sovereign AI model
proton.me2026-09-18about the Apertus familyHN 4 pts · 1 comments671 words
Proton partners with Apertus researchers to integrate the Apertus 1.5 model into Lumo for user feedback. Tens of millions of Lumo users can now voluntarily contribute feedback to improve the open model.generated, cited
Apertus 1.5 out – Latest version of Switzerland's open model with 70B version
cscs.ch2026-07-24about the Apertus familyHN 7 pts · 2 comments684 words
On July 24 2026, ETH Zurich, EPFL, and CSCS released Apertus 1.5, a fully open language model trained on the Alps supercomputer that now understands images and audio alongside text. The project also introduced Apertus Mini, a suite of 16 compact models demonstrating how distillation and quantization can deliver efficient, high-quality models for resource-constrained deployments. Released under the Apache 2.0 open-source license, the model suite is available on Hugging Face and documented on https://www.apertus-ai.org/.written, reviewed, cited
  • Apertus 1.5 incorporates feedback from the first release and adds several new capabilities.
  • The Basel-based outlet Bajour uses the model to trace politicians' positions over time and identify inconsistencies.
  • CSCS is launching an inference service for Apertus 1.5 for all CSCS users and the Swiss academic community.
Swiss public broadcaster to block AI scrapers – except Swiss model Apertus
nzz.ch2026-08-12about the Apertus familyHN 3 pts · 0 comments1,702 words
SRG CEO Susanne Wille blocks uncontrolled data outflow by rejecting training bots and crawlers to negotiate better terms with tech platforms. The SRG plans to provide its archive to the Swiss AI model Apertus, developed with ETH Zurich, to maintain data sovereignty and transparency. While the SRG must cut 900 Stellen to meet a political mandate, the organization aims to strengthen the Swiss location against global tech firms.written, reviewed, cited
Apertus – Open-Source Camera
apertus.org2023-11-26HN front pageabout the Apertus familyHN 184 pts · 34 comments
Apertus: Democratizing Open and Compliant LLMs for ...
arxiv.orgauthorityabout the Apertus family
Apertus 8B Instruct - Intelligence, Performance & Price Analysis | Artificial Analysis
artificialanalysis.aiauthority886 words
Apertus 8B and 70B – a new open multilingual LLM from Switzerland
actu.epfl.ch2025-09-02HN 71 pts · 4 comments1,246 words
Apertus: An open, transparent, multilingual language model
ethz.ch2025-09-02about the Apertus familyHN 16 pts · 4 comments1,340 words
All coverage tracked for this model (23) — append-only, nothing is ever removed
TitleSourcePublishedFirst seenSignals
Apertus: Democratizing Open and Compliant LLMs for ...arxiv.org—2026-10-23exa
Public Summary of Training Content for General-Purpose AI modelsaial.ie—2026-10-23exa
Apertus-8B-Instruct-2509 Benchmark Results · The Aggregatetheaggregate.ai—2026-10-23exa
[Literature Review] Apertus: Democratizing Open and Compliant LLMs for Global Language Environmentsthemoonlight.io—2026-10-23exa
Apertus 8B Instruct - Intelligence, Performance & Price Analysis | Artificial Analysisartificialanalysis.ai—2026-10-23exa
Proton partners with Apertus, Switzerland's sovereign AI modelproton.me2026-09-182026-10-07hn
Swiss public broadcaster to block AI scrapers – except Swiss model Apertusnzz.ch2026-08-122026-10-07hn
Apertus 1.5, Swiss open-weight, open-source, open training data modelapertus-ai.org2026-07-252026-10-07hn
Apertus 1.5 out – Latest version of Switzerland's open model with 70B versioncscs.ch2026-07-242026-10-07hn
Apertus – Open Foundation Model for Sovereign AIapertvs.ai2026-06-212026-10-07hn
Apertus: Democratizing Open and Compliant LLMs for Global Language Environmentsarxiv.org2025-09-222026-10-07hn
Apertus: An open, transparent, multilingual language modelrobohub.org2025-09-162026-10-07hn
Apertus, the most powerful open-source LLM released by a public institutionpublicai.co2025-09-052026-10-07hn
LLM Benchmark Evaluation - Apertus-8B | DS-NLP Labblog.nlp-lab.ai2025-09-052026-10-23exa
Apertus includes many languages that have so far been underrepresented in LLMsswiss-ai.org2025-09-022026-10-07hn
Apertus: A fully open, transparent, multilingual language modelswisscom.ch2025-09-022026-10-07hn
Apertus 8B and 70B – a new open multilingual LLM from Switzerlandactu.epfl.ch2025-09-022026-10-07hn
Apertus: An open, transparent, multilingual language modelethz.ch2025-09-022026-10-07hn
Apertus 8B Instruct benchmark scores — Benchmark Atlasatlas.kevinhu.io2025-09-022026-10-23exa
Apertus – Open-Source Cameraapertus.org2023-11-262026-10-07hn
Apertus: Open Source Cinemaapertus.org2017-09-242026-10-07hn
Apertus, an Open Source Cinema Camerablog.makezine.com2012-02-122026-10-07hn
Apertus - open source cinema Newsnews.ycombinator.com2011-10-292026-10-07hn

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

  • Llama-3.1-8B-Instruct · 8.03B
    same size band (M), same task (text-generation), more providers
  • Qwen3-8B · 8.19B
    same size band (M), same task (text-generation), more providers, permissive license
  • Meta-Llama-3-8B-Instruct · 8.03B
    same size band (M), same task (text-generation)

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:38 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 15:08 UTC · Last verified 2026-10-25

Follow Apertus-8B-Instruct-2509

Watch this model

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