The AI Now report that @F_Kaltheuner wrote dives into this exact question..
RESEARCH
-

Mistral AI releases Mistral Small 3
By
–
Mistral AI announced Mistral Small 3, a latency-optimized 24B-parameter model released under the Apache 2.0 license.
-

LLM Agents Production: Architectures Challenges Best Practices
By
–
LLM Agents in Production: Architectures, Challenges, and Best Practices – ZenML Blog https://
bit.ly/4gGcPtW
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
AI Revolution in R&D: 10X Productivity Gains Explained
By
–
Don't miss my latest #podcast: The AI Revolution In R&D And Innovation Discover how AI is transforming scientific discovery, with some organizations already seeing mind-blowing 10X productivity gains in their R&D departments – this is a revolution you won't want to miss! I'm
-
New 24B Open-Source AI Model Achieves 81% MMLU Performance
By
–
A new model to hasten AI progress. 24B, 81% MMLU, no RL for now! We're super excited to see the latest development in international open-source AI (kudos to Deepseek!), and cannot wait to bring new contributions to it. We're renewing our commitment to using Apache licenses. AI
-

Following AI Model News Updates and Developments
By
–
Watching AI model news be like https://t.co/IY1wLMoItR pic.twitter.com/lydJ3gpm77
— Paul Roetzer (@paulroetzer) 30 janvier 2025Watching AI model news be like
-
Small 3: New Efficient 24B Open Source Language Model
By
–
Introducing Small 3, our most efficient and versatile model yet! Pre-trained and instructed version, Apache 2.0, 24B, 81% MMLU, 150 tok/s. No synthetic data so great base for anything reasoning – happy building!
-
RAG Technology Evolution and Modern AI Applications
By
–
Read more https://
open.substack.com/pub/louisbouch
ard/p/rag-is-old-news?r=25qlky&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true
… -

Evolution of LLM Context Windows and RAG Technology
By
–
In the early days of LLMs, context windows, which is what we send them as text, were small, often capped at just 4,000 tokens (or 3,000 words), making it impossible to load all relevant context. This limitation gave rise to approaches like Retrieval-Augmented Generation (RAG) in
-
The AI Revolution In R&D And Innovation
By
–
The AI Revolution In R&D And Innovation https://
x.com/i/broadcasts/1
ynJODMgQrQxR
…