US Researchers Launch A DeepSeek Competitor

A small team of researchers at Stanford and Washington Universities have created an advanced and very significant AI reasoning model, named s1, for an incredibly low cost of under $50. 

This is highly significant in an industry where developing similar models takes many millions of dollars in resource and infrastructure costs at a time of  growing competition in the AI reasoning field.

For the purpose of comparion, Chinese startup DeepSeek recently made a big impact with its own reasoning model, R1, which the company claims to have been developed for just $6 million.

The s1 model can complete complex reasoning tasks, and has performed in similar ways to OpenAI’s o1 and DeepSeek’s R1 with maths and coding. However, critics are questioning the accuracy of DeepSeek’s claims, and also expressed their concerns regarding the safety and security of its models.

Low Cost Of s1’s Development

This process involves training s1 to mimic the reasoning abilities of an existing AI model, specifically, Google’s Gemini 2.0 Flash Thinking experimental model. By using a curated dataset of 1,000 questions and answers, paired with reasoning traces from the Gemini model, s1 learned how to arrive at accurate solutions in a fraction of the time and cost compared to traditional methods.

According to the researchers, training s1 took just 26 minutes using 16 Nvidia H100 GPUs, costing just $20 in total.

The researchers used what they call Supervised Fine-Tuning (SFT), a method that involves guiding the model with explicit instructions to accelerate the learning process. One particularly interesting development during s1’s creation was the introduction of a “wait” instruction, which helped improve its accuracy. By incorporating pauses into the model’s reasoning process, the researchers found that s1 was able to double-check its responses, often correcting errors and leading to more accurate conclusions.

The researchers behind s1 hope their work will drive open innovation, making powerful reasoning models more accessible to the global community and accelerating advancements in AI technology for the benefit of society. 

However, a higher level of investment may still be necessary to push the envelope of AI innovation. 

The shrort-cut methods used by s1 and R1 (sometimes referred to as distillation) are demonstrably a good method for cheaply re-creating an AI model’s capabilities, but they don’t create new AI models vastly better than what is already available.

arXiv   |    I-HLS    |   Interesting Engineering     |  Tech Xplore   |  Mashable  | Tech Crunch   |   Yahoo

Image: Igor Kutyaev

You Might Also Read: 

A History Of Artificial Intelligence: Its Current & Future Development:


If you like this website and use the comprehensive 6,500-plus service supplier Directory, you can get unrestricted access, including the exclusive in-depth Directors Report series, by signing up for a Premium Subscription.

  • Individual £5 per month or £50 per year. Sign Up
  • Multi-User, Corporate & Library Accounts Available on Request

Cyber Security Intelligence: Captured Organised & Accessible


 

« Thai Police Arrest Russian Hackers
Business Interruption Is The #1 Cyber Risk »

CyberSecurity Jobsite
Check Point

Directory of Suppliers

The PC Support Group

The PC Support Group

A partnership with The PC Support Group delivers improved productivity, reduced costs and protects your business through exceptional IT, telecoms and cybersecurity services.

DigitalStakeout

DigitalStakeout

DigitalStakeout enables cyber security professionals to reduce cyber risk to their organization with proactive security solutions, providing immediate improvement in security posture and ROI.

ZenGRC

ZenGRC

ZenGRC (formerly Reciprocity) is a leader in the GRC SaaS landscape, offering robust and intuitive products designed to make compliance straightforward and efficient.

Practice Labs

Practice Labs

Practice Labs is an IT competency hub, where live-lab environments give access to real equipment for hands-on practice of essential cybersecurity skills.

North Infosec Testing (North IT)

North Infosec Testing (North IT)

North IT (North Infosec Testing) are an award-winning provider of web, software, and application penetration testing.

European Cybercrime Training and Education Group (ECTEG)

European Cybercrime Training and Education Group (ECTEG)

The primary aim of ECTEG is to enhance the coordination of cybercrime training, by identifying opportunities to build the capacity of countries to combat cybercrime

BeOne Development

BeOne Development

BeOne Development provide innovative training and learning solutions for information security and compliance.

Protocol Policy Systems

Protocol Policy Systems

Protocol Policy Systems specialise in IT policy deployment and management systems that deliver compliance and secure computing environments.

SISSDEN

SISSDEN

SISSDEN will improve cybersecurity through the development of increased awareness and the effective sharing of actionable threat information.

CS3STHLM

CS3STHLM

CS3STHLM is the Stockholm international summit on Cyber Security in SCADA and Industrial Control Systems.

Evanston Technology Partners (ETP)

Evanston Technology Partners (ETP)

ETP provides services and solutions to enable and transform businesses in the areas of cybersecurity, data protection, and efficient operations practices.

GitGuardian

GitGuardian

Enable developers, ops, security and compliance professionals to enforce security policies across public and private code, and other data sources as well

Inceptus

Inceptus

Inceptus is a next generation Managed Security Service Provider (MSSP). We are dedicated to keeping our customers safe, secure and protected while doing business on the Internet.

iSecurity Consulting

iSecurity Consulting

iSecurity delivers a complete lifecycle of digital protection services across the globe for public and private sector clients.

Bechtle

Bechtle

Bechtle is one of Europe’s leading IT service providers offering a blend of direct IT product sales and extensive systems integration services.

AlJammaz Technologies

AlJammaz Technologies

AlJammaz Technologies is the leading Technology Value-Added Distributor, which distributes advanced technology products, solutions and services in area including networking and cybersecurity.

RedHunt Labs

RedHunt Labs

RedHunt Labs is a premier Cybersecurity Solutions provider, offering Attack Surface Management solution 'NVADR' and Penetration Testing services.

Aura

Aura

Aura is a mission driven technology company dedicated to creating a safer internet for everyone. We’re making comprehensive digital security that's simple to understand and easy to use.

iSTORM

iSTORM

iStorm specialise in supporting organisations who require a range of Privacy, Security and Penetration testing related services.

DataPatrol

DataPatrol

DataPatrol is a software company, specialized in providing Security and Privacy of company’s data and information in an evolved way.

Liverton Security

Liverton Security

Liverton Security is a New Zealand-owned cyber security provider offering consultancy and security-related products to government and commercial customers throughout New Zealand.