Hey HN,

Henry from Cactus here!

We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2.

The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges 300-700 on sub-$200 phones such as the Samsung A-Series.

On the tool call and mobile device use benchmarks, Needle 2 trades wins with closest small models like LFM2.5 230M and Apple Foundation Model, at 5x to 70x smaller, both at f16 vs Needle 2 at 2bit. Needle is based on Simple Attention Networks from our paper (https://arxiv.org/abs/2607.18363).

Edge AI has lately meant Macs and PCs, but that is just 1.5 billion of over 21 billion connected IoT devices in the world today, and in emerging markets most phones ship under $200, no NPU, cheap GPUs. These include budget phones, Raspberry Pis, microcontrollers, wearables, small robots like Reachy Mini, and connected home devices.

A conventional transformer of Needle's width and depth spends 164 MFLOPs per token, and even one squeezed down to Needle's parameter count spends 87, Needle spends 70. Even on a high-end phone, an always-on assistant lives inside a power budget; every MFLOP is milliwatt-hours, and Needle spends 7x to 85x fewer of them per token than the smallest performant LLMs. More about the architecture in the link.

When we structure intelligence for consumer devices as functions with typed parameters, the only hard part is mapping a messy sentence onto them; which function, with which values. Our research found that when framed that way, the problem needs no world knowledge and no open-ended prose, which is why 45M parameters suffice.

Needle 2 expands to structured extraction where the schema can be passed in-place of tools and the model returns structured output. You can use Needle as a text-classification model with an enum field, as a summarization model by providing a schema that extracts key fields, everything but free-range decode.

Every product has its own tool vocabulary and fine-tuning needle helps it achieve frontier-level performance on custom tasks, so using the python package (https://github.com/cactus-compute/needle), Needle can be fine-tuned Needle on a Mac/PC in minutes to a few hours, with automated data-generation pipeline, just pass a couple samples.

Nonetheless, every response carries a learned confidence score based our Cactus Hybrid technique. If above your threshold, act, below it, escalate to the cloud or bigger model. Combining Needle 2 with a private DeepSeek-v4-Flash deployment works particularly well for enterprise-level tasks at barely any cost, we can help with this setup.

We have put a lot of thoughts into Needle 2 but might still be missing quite a lot, please use the playground in the provided link to test Needle and share your thoughts, always appreciated!

Bernie Sanders Writes Letter to Sam Altman, Dario Amodei and Mark Zuckerberg [pdf]

NCAA and FBI partner up to help athletes avoid being sexually exploited online

It's Official: Flock Is for Stalking

Learning more about Claude's mathematical capabilities

An Underscore Sent an Innocent Man to Jail

Inverse Differential Equations

Learning from Historical Mistakes

Former acting IRS head warns of tax compliance 'race to the bottom'

Cylic Trait Implementations: Motivation

Why play is at the core of a meaningful life

Hi HN Rune – A software intelligence runtime for AI coding assistants

Who Should Pay for Source Code Availability?

We're releasing a new model (GPT-5.6-Cyber)

Frontier Intelligence for Cybersecurity Teams

AI Boot Camps Surge as Workers Race to Learn New Skills

The Case for the Everettian Multiverse

When coding gets cheap, engineering gets harder

Cloudflare Tools Secure Autonomous AI Payments

Following Epic loss, Google hosts rival app stores in the Play Store

CEO Just Fired 500 People Because He Says Zillow Is More Efficient Without Them

Ask HN: How do you learn with LLMs?

Lindy Effect

Bayesianity

Configuration Is Not Customization

OpenAI launches GPT-5.6-Cyber with fewer refusals for exploit research

Xirp, a vendor-neutral agentic development environment by Spotify

Show HN: Needle2: 14MB agentic LLM for phones, wearables, smart home and robots

Hyper-Exponential Breakthroughs

Pave the Road

Show HN: Build on textlog

Show HN: Git-chain – Simplify branch dependency management in Git

Comments