Aleph Alpha (@Aleph__Alpha) header
Aleph Alpha (@Aleph__Alpha) profile picture

Our Specialized Sovereign Large Language Model for Mission Critical Environments is here: https://teapawt.org/t.co/261IusY3RI

Heidelberg, Germany
Joined December 2018
Aleph Alpha retweeted
Dear @Aleph__Alpha team - thank you for making Kolibri-1 open. We care deeply about sovereign AI, and launching it on German Unity Day makes today feel especially fitting. As a small gesture of support, we’ve hosted and made Kolibri-1 free for anyone to try for the next few days. No GPU. No setup. Just try it. tesseracted.com/kolibri-1-ch… Open Models move us all forward. We’re rooting for you. 🇩🇪 Tesseracted Labs GmbH. tesseracted.com #alephalpha #openweightmodels
23
43
11
429
52,909
Aleph Alpha retweeted
germany just dropped a sovereign open weight model kolibri by @Aleph__Alpha runs 3.5b of its 78b parameters per word and its math is kinda ridiculous: 96.9% on aime beats every mixture-of-experts model they tested, even 3x bigger ones. only a dense model doing 8x the work wins anyone can run it on their own servers, it thinks in german, and in their evals it tops every open model its size in english and german models read text in chunks called tokens, and i ran kolibri's chunker (its tokenizer) on the german constitution: it needed 15% fewer tokens than gpt-5's for the same text. "bundesverfassungsgericht" is 6 tokens for gpt-5 and 2 for kolibri. fewer tokens means cheaper, faster german, and more of it fits in what the model can read at once how it works, simply: 1. every layer has 384 tiny specialists, and a router sends each word to 6 of them. so it thinks like a 3.5b model, but it needs the memory of a 78b one: about 78 gb, which means 2 big nvidia gpus (h100s) or 1 h200 2. most layers only look at the last 512 tokens, and every 5th layer looks at everything. that's how it can read 1 million tokens (a few thick books) without it costing a fortune 3. it reasons in german. their team found that a little german reasoning data is worse than none: the model's german thoughts go in circles and never finish. so they made about 800k german reasoning examples and gave it a lot 4. it's trained to say "i don't know". they play a game with it where parts of the documents are hidden, sometimes to help it and sometimes to hide the evidence, and it has to tell which. when it didn't know an answer it admitted it 44% of the time. qwen3.5 did 11% where it's weaker: answering from memory, using tools over a long back and forth, and coding agents, where qwen models are ahead. and to run it you need aleph alpha's add-on for vllm, a popular open source server for running models if you have german documents and need to keep them on your own hardware, this is a big deal. huge congrats to everyone at aleph alpha, my good friend @MichaelLHofmann included!! i wrote up how it works, the benchmarks, how to run it and when to use it: tej.as/blog/aleph-alpha-koli…
26
60
11
497
33,321
Aleph Alpha retweeted
Hot off the press: We are becoming the first transatlantic sovereign AI solution together with our partner @cohere. More talent, more compute, and more innovation power to offer trustworthy AI at the security level that governments and enterprises need – across the globe.
3
15
4
88
5,740
Aleph Alpha retweeted
🚀 Sovereign AI for the world. Cohere & Aleph Alpha form transatlantic AI powerhouse anchored in Canada & Germany! Combining our global scale with European R&D excellence to build sovereign, enterprise-grade AI. Security, privacy & trust for businesses & governments worldwide. #SovereignAI #AIPartnership Learn more: businesswire.com/news/home/2… Image from left to right: Rolf Schumann, Schwarz Digits, Samuel Weinbach, Aleph Alpha, Aidan Gomez, Cohere, Minister Solomon, Canada, Minister Wildberger, Germany
9
27
19
189
122,359
„Europa muss sich bei #genAI nicht verstecken. Im Gegenteil: Wir spielen vorn mit. Nun kommt es auf kluge Regulierung an, die international abgestimmt sein muss, wie #G7 #CoC, den 🇩🇪 mit 🇺🇸 voranbringen konnte“, so @Wissing im Pressegespräch mit @JonasAndrulis von @Aleph__Alpha
19
9
3
37
10,506
Aleph Alpha retweeted
Digitalisierung der öffentlichen Verwaltung mittels KI?🧠 Ganz genau. Wir freuen uns, mit @Aleph__Alpha eine strategische Partnerschaft für den öffentlichen Sektor geschlossen zu haben, um die Digitalisierung auf das nächste Level zu heben. Alle Details: di-ri.co/A9Pa5
1
3
11
2,829
Are you at #IFABerlin2023 - let's chat about generative AI beyond chatbots and copywriting.
1
1
16
2,753
Faszinierende Gespräche über KI mit drei der führenden Experten aus Deutschland: @nettwerkerin Prof. Katharina Zweig, Christopher Kränzler ⁦@lengooTweets⁩, Jonas Andrulis #AlephAlpha. Auch im Vergleich zu ChatGPT hat Deutschland einiges zu bieten. KI mit Transparenz…
489
19
31
331
213,986
#KI ist mehr als Technologie – sie ist der Weg in eine bessere Zukunft. Sie schafft Chancen und Wachstum für Deutschland. Nur gemeinsam gestalten wir eine vertrauenswürdige KI 'Made in Europe', so Volker @wissing zur Partnerschaft zwischen @Aleph__Alpha & #IPAI in Heilbronn.
1
2
1
11
3,629
Aleph Alpha retweeted
Ein Gespräch über apokalyptische Ängste, den Konflikt zwischen Europa und dem Silicon Valley und Witze über Möwen: @Aleph__Alpha-Mitgründer @JonasAndrulis im Interview #ki #chatgpt #AIact #SZPlus sz.de/1.6034862?utm_source=T…
2
3
2,388
Regulierung von #KI ja, aber mit Augenmaß. Wie das geht, haben #o2 Telefónica CFO Markus Rolle, @DanielaKluckert, @alena_buyx, @BirgitSippelMEP & Tobias Haar/@Aleph__Alpha bei der #DataDebates diskutiert. Debatte verpasst? Jetzt ansehen: tinyurl.com/cdmupsdc
1
4
1
13
173,843
Aleph Alpha retweeted
Heute sind wir zum Digitallunch zu Gast beim @aicampusberlin. Jetzt im Gespräch mit @Hansi_Stu @Aleph__Alpha und @hendrik_jandel @Merantix zu Künstlicher Intelligenz zwischen Chancen, Potentialen, Risiken und Nebenwirkungen. #ArtificialIntelligence #AIAct
3
10
2,052
A lot of smart people are thinking about how to best regulate AI. From a European perspective the first question should be: Hoe can we not just regulate but build the future? Our contribution can be much more than the cookie banner.
1
5
23
4,332
Trust me, I'm an AI system? Erklärbarkeit und #Vertrauen sind Knackpunkt generativer KI und Thema im #AIAct. Das #KI-System von @Aleph__Alpha hat eine neue Control-Version erhalten: Wir haben bei @JonasAndrulis und @TimAndreThomas nachgehorcht 🇩🇪/🇬🇧 heise.de/news/Erklaer-KI-Ale…
3
9
1,592
Join us at #hubberlin23 as we dive into the world of #AI and explore its transformative potential!🌟@JonasAndrulis, Founder & CEO of @Aleph__Alpha, will hold a keynote on day 1! For more updates, visit our newsroom 👉 hub.berlin/news #hubberlin #hubspeakers
2
2
1,340
Aleph Alpha retweeted
"Maschinen haben nur gelernt, Sprache fortzuschreiben. [...] Wenn gewisses Wissen nicht stark genug da ist, wird halt irgendwas erzählt. [...] Wir arbeiten daran, dass die Ergebnisse zurückgeführt werden können auf die Quellen." (@JonasAndrulis, @Aleph__Alpha) #OMR23 #AI #KI
2
1
15
4,152
Brilliant keynote presentation by @JonasAndrulis CEO @Aleph__Alpha - a leader of artifical general intelligence which can actually evidence that the answers given by their models are correct. Available in 5 languages 👏👍🏽 #fintech #AI #IFGS2023
11
18
1,475
New @officially_omr podcast: @JonasAndrulis of @Aleph__Alpha discusses successes, challenges & latest funding rounds. 🇩🇪Don't miss a conversation on how a European counterpart to OpenAI & Microsoft poses high value + big implications. @AndreRetterath 🎧 https://teapawt.org/t.co/v4Wus7djQT
1
5
2
13
2,757
German AI startup @Aleph__Alpha claims it will have a 300-billion parameter Large Language Model "Luminous-World" trained this year for “highly complex and critical applications” (GPT-3 has 175 billion parameters)
16
48
7
342
88,015