Excited to be (finally) heading to the PyTorch Conference! I’ll be giving a talk tomorrow at 11:00 AM on “The LLM Landscape 2025”, where I’ll discuss the key components behind this year’s most prominent open-weight LLMs, and highlight a few architectural developments that go
OPEN SOURCE
-
Qwen3 Small Models: Faster and Cheaper Alternative to GPT
By
–
Interesting that they mentioned faster & cheaper compared to OpenAI’s latest models not “customizable”. That makes me think they are specifically referring to gpt-oss, This in turn means they are using the small, dense Qwen3 models, maybe 0.6 to 4B range. And this is
-
Training multilingual cultural model improves multimodal benchmark performance
By
–
Thank you so much and appreciate you taking time to check the paper. We trained a model(
https://
huggingface.co/neulab/Cultura
lPangea-7B
…) on the subset of the dataset and were able to improve performance on many multilingual/cultural multimodal benchmarks. Yeahhh, would be great too if you could also -
Stable Audio 2.5 live chat with ComfyUI demonstrates inpainting
By
–
We joined our friends at @ComfyUI for a live chat about what's new with Stable Audio 2.5 🥁
— Stability AI (@StabilityAI) 21 octobre 2025
Check out CJ Carr from our Audio Research team demoing the model’s inpainting capability.
You can watch the full episode here for more insights and examples 👉 https://t.co/430lf6kve3 pic.twitter.com/2A0KHZieQBWe joined our friends at @ComfyUI for a live chat about what's new with Stable Audio 2.5 🥁 Check out CJ Carr from our Audio Research team demoing the model’s inpainting capability. You can watch the full episode here for more insights and examples 👉 bit.ly/43nRa5x
→ View original post on X — @stabilityai, 2025-10-21 19:31 UTC
-
NVIDIA Nemotron: How Open Source AI Models Transform Development
By
–
👀 What happens when AI goes open source?
— NVIDIA AI (@NVIDIAAI) 21 octobre 2025
NVIDIA leaders share how Nemotron, our family of open models, is changing how the world builds, customizes, and trusts AI.
🎧 Listen now: https://t.co/hfwD2tpwCA pic.twitter.com/rc03h8gmhHWhat happens when AI goes open source? NVIDIA leaders share how Nemotron, our family of open models, is changing how the world builds, customizes, and trusts AI. Listen now: https://
nvda.ws/4nj5a86 -
New Tokenization Approach Challenges BPE-Style Methods
By
–
TLDR: surprised that it works, skeptical whether this will replace BPE-style tokenizers, delighted about this refreshing approach
-
Reproducible Builds on Android iOS Apple Changes Required
By
–
We have repro builds on Android, and we do as much as we can on iOS. For any app to be fully reproducible on iOS will require that Apple make some changes (please pressure them to do so).
-
Open-weight PyTorch models security and trustworthiness discussion
By
–
If we are talking about the model itself and not the app, these are open-weight PyTorch models. So unless there’s a backdoor in Hugging Face or the PyTorch runtime, there’s really no way for them to be malicious afaik.
-
DeepSeek OCR Paper and Multimodal LLM Article Resources
By
–
Link to the paper: https://
github.com/deepseek-ai/De
epSeek-OCR/blob/main/DeepSeek_OCR_paper.pdf
… My "Understanding Multimodal LLMs" article with more info on how images are fed to LLMs, how cross-attention works, etc: https://
magazine.sebastianraschka.com/p/understandin
g-multimodal-llms?utm_source=publication-search
… -
Motia: Unified Backend Framework for Building AI Agents
By
–
A unified backend framework to build and scale AI Agents!
— Sumanth (@Sumanth_077) 21 octobre 2025
Motia is a backend framework that brings together agents, APIs, background jobs, streams, and workflows into a single core primitive.
Everything is event driven and observable.
It’s 100% Open Source. pic.twitter.com/InYzshW8ZEA unified backend framework to build and scale AI Agents! Motia is a backend framework that brings together agents, APIs, background jobs, streams, and workflows into a single core primitive. Everything is event driven and observable. It’s 100% Open Source.