Microsoft Unveils BitNet: A Hyper-Efficient AI Model for CPUs

Apr 17, 2025
Microsoft researchers have introduced BitNet b1.58 2B4T, a hyper-efficient AI model that can run on CPUs, including Apple's M2, offering significant computational efficiency.

Microsoft researchers have introduced BitNet b1.58 2B4T, a hyper-efficient AI model designed to run on CPUs, including Apple's M2. This model, described as the largest-scale 1-bit AI model or 'bitnet' to date, is openly available under an MIT license. Bitnets are compressed models that quantize weights into three values: -1, 0, and 1, making them more memory- and computing-efficient than traditional models.

BitNet b1.58 2B4T, with 2 billion parameters, was trained on a dataset of 4 trillion tokens. It reportedly outperforms traditional models of similar sizes, such as Meta's Llama 3.2 1B and Google's Gemma 3 1B, on benchmarks like GSM8K and PIQA. The model is also noted for its speed, operating at twice the speed of other models of its size while using significantly less memory.

However, to achieve optimal performance, BitNet requires Microsoft's custom framework, bitnet.cpp, which currently supports only certain hardware, excluding GPUs. This limitation highlights a potential challenge in broader adoption, as GPUs are prevalent in AI infrastructure.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Daily AI Brief.

Also, consider following us on social media:

AI Brief AI Brief (X)
Free newsletter

Daily AI Brief

Daily report covering major AI developments and industry news, with both top stories and complete market updates

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.