← Back to BLACKWIRE CIPHER BUREAU AI SURVEILLANCE A diagram of the Needle model architecture, showing its retrieval-and-assembly process

The Needle model is designed to facilitate retrieval-and-assembly, making it an ideal solution for developers looking to create AI-powered tools. Cactus Compute's investigation into agentic models led to the development of this groundbreaking technology.

CACTUS COMPUTE UNLEASHES NEEDLE: 26M PARAMETER MODEL REVOLUTIONIZES TOOL CALLING ON BUDGET DEVICES

_In a groundbreaking move, Cactus Compute has open-sourced Needle, a 26M parameter function-calling model that runs on consumer devices. This development has significant implications for the field of artificial intelligence and could potentially disrupt the status quo. The model's ability to run at 6000 tokens per second prefill and 1200 tokens per second decode on budget phones makes it an attractive solution for developers._

By CIPHER Bureau - BLACKWIRE | May 13, 2026, 05:00 CET | AI, surveillance, cybersecurity, artificial intelligence, machine learning

Cactus Compute has made a major breakthrough in the field of artificial intelligence with the release of Needle, a 26M parameter function-calling model. The model is designed to run on consumer devices, making it an attractive solution for developers looking to create AI-powered tools. With its ability to run at 6000 tokens per second prefill and 1200 tokens per second decode, Needle has the potential to revolutionize the field of artificial intelligence.

The Problem with Current Models

Current agentic models are often too large and resource-intensive, making them impractical for use on budget devices. This limitation has hindered the development of AI-powered tools that can run on consumer hardware. Cactus Compute's investigation into this issue led to the realization that massive models are overkill for tool calling, which is fundamentally a retrieval-and-assembly process. By distilling the Gemini tool calling into a 26M parameter model, the company has made a significant breakthrough.

The Technology Behind Needle

Needle is built on the principles of function-calling and tool use, allowing it to run efficiently on consumer devices. The model's architecture is designed to facilitate retrieval-and-assembly, making it an ideal solution for developers looking to create AI-powered tools. With its ability to run at 6000 tokens per second prefill and 1200 tokens per second decode, Needle has the potential to revolutionize the field of artificial intelligence.

We were always frustrated by the little effort made towards building agentic models that run on budget phones, so we conducted investigations that led to an observation: agentic experiences are built upon tool calling, and massive models are overkill for it.

Implications and Potential Applications

The release of Needle has significant implications for the field of artificial intelligence. The model's ability to run on budget devices makes it an attractive solution for developers looking to create AI-powered tools for a wide range of applications. From surveillance and security to healthcare and education, the potential uses of Needle are vast. As the model continues to evolve and improve, it is likely to have a major impact on the way we interact with technology.

The Future of AI-Powered Tools

The release of Needle is a significant step forward in the development of AI-powered tools. As the model continues to evolve and improve, it is likely to enable the creation of a wide range of new tools and applications. From intelligent assistants to autonomous vehicles, the potential uses of Needle are vast. As the field of artificial intelligence continues to advance, it is likely that we will see a major shift in the way we interact with technology.

Sources: Cactus Compute, Hacker News, GitHub