Apertus retweeted
We have achieved well over 100x better training efficiency for the next Apertus model!
👇
Apertus retweeted
Another alarming finding is the big gap between closed and open models when it comes to understanding less represented cultures - our Apertus team now has a reliable way to help us make our next v2 model more multicultural 🤗
The 3rd #OpenSourceLLM Builders Summit brought open model developers for sessions in Zurich on relevance, sustainability of open model development: swiss-ai.org/3rd-open-source…
#Apertus #AISingapore #LAION #OpenAthena #Marin #K2Horizon #OpenEuroLLM #CurrentAI #Nemotron #CSCS #EuroHPC
1
5
171
Apertus retweeted
In v1.5 we had to build a foundation of our post-training practices, starting from data - cleaning, filtering, decontamination are the standard, but in some areas data is very scarce
We have been cooking some open pipelines for underrated domains to scale data, more updates soon!
Apertus retweeted
Apertus 1.5 Technical Report is out!
134 pages of fully open multimodal reasoning model development from Swiss AI Initiative (ETH Zurich/EPFL) with data recipe, training experiments and many interesting insights 🇨🇭
Thread with my personal coverage ⬇️
Get started with #Apertus #LLM - now easier as the fully open model is deployed into AI systems by top Swiss tech teams 🇨🇭 Learn about the first inference providers working with research, contributing to key #opensource components in our blog apertus-ai.org/articles/2026…
3
269
The Apertus team, a group of researchers at EPFL, ETH Zurich, and the Swiss National Supercomputing Centre (CSCS), partners with Swiss technology company Proton to make Apertus 1.5 available in Lumo, Proton’s AI assistant.
actu.epfl.ch/news/apertus-pa…
Apertus retweeted
Das Schweizer KI-Modell Apertus nach einem Jahr: Was funktioniert und was nicht? Wir haben bei Fachleuten aus der Branche nachgefragt.
swissinfo.ch/ger/schweizer-k…
Apertus retweeted
A un anno dal lancio, Apertus ha superato i 4 milioni di download e trovato applicazioni concrete. La nuova versione presentata a luglio riduce il divario con i concorrenti, ma competere nel settore dell’IA resta una sfida per il modello svizzero. buff.ly/IYzT3uW
EPFL, @ETH_en & the Swiss National Supercomputing Centre (CSCS) release #Apertus 1.5: a fully open, multimodal LLM. It's now understanding text, images & audio, powering sovereign AI solutions for research, public administration and industry. actu.epfl.ch/news/apertus-15…
Available now! #Apertus 1.5 — 8B & 70B models with multimodal text/image/audio input, 4x longer context window, optional thinking mode, improved tool use, all built transparently and responsibly. 3.2M+ downloads 🚀 #OpenSourceLLM Details+Links+Open roles: apertus-ai.org/articles/2026…
Apertus retweeted
Hey 👋 Apertus 1.5 is now available to the public on publicai.co! 😍
Chat with it here: chat.publicai.co/
Try out the image capabilities! I'd be curious what people think before sharing detailed benchmark results :)
Apertus retweeted
Hot open-source summer season has gotten to Switzerland 🇨🇭⛰️
Our team at Swiss AI Initiative is happy to release Apertus 1.5 - a multimodal and reasoning update to our first v1 version!
The new version builds a foundation for future regular open model development at scale🧵
Apertus retweeted
Today we are releasing Apertus 1.5 8B and 70B, a continued pretraining of our Apertus 1.0 models, which have passed 3.2M downloads in the community. 🧵
Apertus retweeted
After escaping European heat with 30+ hrs delay and missing luggage I am finally in San Diego attending @aclmeeting! 🇺🇸🌴#ACL2026
I will be presenting our work on the multilingual benchmark translation from @INSAITinstitute 🇺🇦🇧🇬
📍Poster Session E, July 6th 11.00-12.30
Also check out @apertusllm poster, presented by the greatest @agromanou, to know more about the biggest fully open multilingual models 🇨🇭
📍Poster Session F, July 7th 9.00-10.30
Three new model weights: 0.5B, 1.5B, 4B are available using new quantization and distillation techniques. Download #Apertus 1.1, read the ICML workshop report, try a new demo - all just a tap away in our latest blog post: apertus-ai.org/articles/2026…
Apertus retweeted
Apertus Mini is now running entirely in your browser 🇨🇭
80+ tps for the 1.5B, 60+ tps for the 4B (on my M3). Fully client-side via Transformers.js + ONNX + WebGPU.
Congratulations to each of the 102 authors of the #Apertus v1 report, officially published today ahead of the presentation at #ACL2026NLP in July. Follow @aclmeeting for updates aclanthology.org/2026.acl-lo…













