Researchers open source Sky-T1, a ‘reasoning’ AI model that can be trained for less than $450

Date:

Share post:


So-called reasoning AI models are becoming easier — and cheaper — to develop.

On Friday, NovaSky, a team of researchers based out of UC Berkeley’s Sky Computing Lab, released Sky-T1-32B-Preview, a reasoning model that’s competitive with an earlier version of OpenAI’s o1 on a number of key benchmarks. Sky-T1 appears to be the first truly open source reasoning model in the sense that it can be replicated from scratch; the team released the data set they used to train it as well as the necessary training code.

“Remarkably, Sky-T1-32B-Preview was trained for less than $450,” the team wrote in a blog post, “demonstrating that it is possible to replicate high-level reasoning capabilities affordably and efficiently.”

$450 might not sound that affordable. But it wasn’t long ago that the price tag for training a model with comparable performance often ranged in the millions of dollars. Synthetic training data, or training data generated by other models, has helped drive costs down. Palmyra X 004, a model recently released by AI company Writer, trained almost entirely on synthetic data, reportedly cost just $700,000 to develop.

Unlike most AI, reasoning models effectively fact-check themselves, which helps them to avoid some of the pitfalls that normally trip up models. Reasoning models take a little longer — usually seconds to minutes longer — to arrive at solutions compared to a typical non-reasoning model. The upside is, they tend to be more reliable in domains such as physics, science, and mathematics.

The NovaSky team says it used another reasoning model, Alibaba’s QwQ-32B-Preview, to generate the initial training data for Sky-T1, then “curated” the data mixture and leveraged OpenAI’s GPT-4o-mini to refactor the data into a more workable format. Training the 32-billion-parameter Sky-T1 took about 19 hours using a rack of 8 Nvidia H100 GPUs. (Parameters roughly correspond to a model’s problem-solving skills.)

According to the NovaSky team, Sky-T1 performs better than an early preview version of o1 on MATH500, a collection of “competition-level” math challenges. The model also beats the preview of o1 on a set of difficult problems from LiveCodeBench, a coding evaluation.

However, Sky-T1 falls short of the o1 preview on GPQA-Diamond, which contains physics, biology, and chemistry-related questions a PhD graduate would be expected to know.

Also important to note is that OpenAI’s GA release of o1 is a stronger model than the preview version of o1, and that OpenAI is expected to release an even better-performing reasoning model, o3, in the weeks ahead.

But the NovaSky team says that Sky-T1 only marks the start of their journey to develop open source models with advanced reasoning capabilities.

“Moving forward, we will focus on developing more efficient models that maintain strong reasoning performance and exploring advanced techniques that further enhance the models’ efficiency and accuracy at test time,” the team wrote in the post. “Stay tuned as we make progress on these exciting initiatives.”



Source link

Lisa Holden
Lisa Holden
Lisa Holden is a news writer for LinkDaddy News. She writes health, sport, tech, and more. Some of her favorite topics include the latest trends in fitness and wellness, the best ways to use technology to improve your life, and the latest developments in medical research.

Recent posts

Related articles

Apple board opposes proposal to abolish DEI programs

Apple’s board of directors has come out in opposition to a proposal seeking to end the company’s...

What’s behind ballooning video game budgets?

Why does it cost some companies hundreds of millions of dollars to develop a popular video game? A...

Everything you missed at CES 2025

Welcome back to Week in Review. I missed you! This week, we’re diving into all the gadgets...

I got soaked driving the Arc Sport electric boat

I did not go to CES 2025 with the goal of getting drenched by Lake Mead’s chilly...

Matt Mullenweg deactivates WordPress contributor accounts over alleged fork plans

Automattic CEO and WordPress co-creator Matt Mullenweg has deactivated the accounts of several WordPress.org community members, two of...

We talked to the guy who was stuck in a Waymo robotaxi on a dizzying loop

A month ago, a video circulated around social media of a Waymo robotaxi stuck in a roundabout...

The US has a new cybersecurity safety label for smart devices

The White House this week announced a new label for internet-connected devices, the U.S. Cyber Trust Mark,...

The LiberLive C1 stringless instrument isn’t a guitar, and that’s fine

I’m old enough to remember musicians getting angry over Guitar Hero. But they always seemed to be...