.png)
Today, we are announcing 1-bit Bonsai models that bring advanced intelligence to the devices where people actually live and work.
PrismML on Twitter / X
1-bit Bonsai 8B running locally on an M4 Pro (MLX) alongside a standard 16-bit 8B model.Same class of model, very different deployment profile: far lower memory use and substantially higher throughput. pic.twitter.com/dxMFSyCZDG— PrismML (@PrismML) March 31, 2026
Bonsai 1-bit WebGPU - a Hugging Face Space by webml-community
Run 1-bit Bonsai LLMs locally in your browser on WebGPU
Extensible Software in the age of LLMs | Jeremy Morrell
Solid core + capability-based sandboxes + LLMs = Users with superpowers
Minions: where local and cloud LLMs meet· Ollama Blog
Avanika Narayan, Dan Biderman, and Sabri Eyuboglu from Christopher Ré's Stanford Hazy Research lab, along with Avner May, Scott Linderman, James Zou, have developed a way to shift a substantial portion of LLM workloads to consumer devices by having small on-device models (such as Llama 3.2 with Ollama) collaborate with larger models in the cloud (such as GPT-4o).
Needle 2 - The 14 MB Agentic LLM for Tiny Devices | Cactus
An open 45M-parameter model for tool calling, device use, and structured extraction. Needle 2 runs as a 14 MB binary in 28 MB of session RAM.
The Ultimate Guide to Fine-Tuning LLMs from Basics to Breakthroughs: An Exhaustive Review of Technologies, Research, Best Practices, Applied Research Challenges and Opportunities (Version 1.0)
Project Glasswing: what Mythos showed us
In recent weeks, we pointed Mythos and other security-focused LLMs at live code across critical parts of our infrastructure. We share what we observed, the models’ strengths and weaknesses, and what the work around them needs to look like before any of it can scale.
Pieces | Infinite Artificial Memory for your Digital Workers and Agents
Pieces is your AI companion that captures live context from browsers to IDEs and collaboration tools, manages snippets and supports multiple llms - all while processing data locally for maximum control.
Introducing SubQ 1.1 Small
Subquadratic is a frontier AI research and infrastructure company building a new class of LLMs.
1-bit Hokusai’s ”The Great Wave” – Hypertalking
5 years ago I started a now completely stalled project (fingers crossed I can figure out how to restart soon) to draw all of Hokusai’s 36 views of Mount Fuji as 1-bit pixel art.
Introducing any-llm: A unified API to access any LLM provider
When it comes to using LLMs, it’s not always a question of which model to use: it’s also a matter of choosing who provides the LLM and where it is deployed. Today, we announce the release of any-llm, a Python library that provides a simple unified interface to access the most popular providers.
distil labs — Replace LLMs with Custom Small Language Models
Train and deploy custom small language models that are faster, cheaper, and just as accurate as LLMs.
Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things
Friday’s big release was Qwen 3.8 27B, an Apache 2 licensed 27B parameter vision-capable LLM from Alibaba’s Qwen research lab. I’ve been looking forward to this one: 27B is an …
Introducing LFM2: The Fastest On-Device Foundation Models on the Market | Liquid AI
Today, we release LFM2, a new class of Liquid Foundation Models (LFMs) that sets a new standard in quality, speed, and memory efficiency for on-device deployment. Built on a hybrid architecture, LFM2 delivers 200% faster decode and prefill performance than Qwen3 and Gemma 3 on CPU. It also significantly outperforms models in each size class on instruction-following and function calling—the core capabilities that make LLMs reliable for building AI agents.
Solving a Million-Step LLM Task with Zero Errors
LLMs have achieved remarkable breakthroughs in reasoning, insights, and tool use, but chaining these abilities into extended processes at the scale of those routinely executed by humans,...
Alex Cheema on Twitter / X
.@karpathy shouted out my work on @exolabs at @ycombinator AI SUS!“we use LLMs similarly to mainframes in the ‘70s - compute is timeshared by having a slice in the batch dimension. models will compress over time, and with this we’ll be able to run more on-device” pic.twitter.com/UKSrquQGAL— Alex Cheema (@alexocheema) June 18, 2025
