U.S. Researchers Build Advanced Reasoning Model For Less Than $50 - 1

Photo by Sebastien Bonneval on Unsplash

U.S. Researchers Build Advanced Reasoning Model For Less Than $50

  • Written by Andrea Miliani Former Tech News Expert
  • Fact-Checked by Justyn Newman Former Lead Cybersecurity Editor
  • Reader’s Comments 1

AI researchers from the University of Washington and Stanford trained an AI reasoning model for less than $50—in cloud computing credits—called s1. The team released a paper, titled s1: Simple test-time scaling , with more details of their methodology this Monday.

In a Rush? Here are the Quick Facts!

  • AI researchers from the University of Washington and Stanford trained an AI reasoning model for less than $50 and shared their research this Monday.
  • They used the distillation technique, a test-time scaling, and a supervised fine-tuning approach, with a 1,000-question dataset.
  • The model s1 performs similarly to DeepSeek R1 and OpenAI o1.

According to TechCrunch, the new model performs similarly to advanced models like DeepSeek ’s R1, or OpenAI’s o1 and is available on GitHub.

To develop the AI model, the researchers applied a process known as distillation—when a larger AI model provides data to a smaller model—getting reasoning capabilities from Google’s Gemini 2.0 Flash Thinking Experimental.

This process is gaining popularity in the AI industry as OpenAI claims that DeepSeek used the process, without authorization, to develop its advanced reasoning model. Researchers from UC Berkeley ’s Sky Computing Lab also recently managed to train a reasoning model for less than $450 with this technique, which is sparking debate in Silicon Valley and anger among large AI companies.

The researchers developing the s1 model also considered a “test-time scaling” approach —by forcing the model to stop and reason more before providing an answer—and performed supervised fine-tuning from a pre-trained model to build its AI reasoning model.

“We develop budget forcing to control test-time compute by forcefully terminating the model’s thinking process or lengthening it by appending ‘Wait’ multiple times to the model’s generation when it tries to end,” states the paper . “This can lead the model to double-check its answer, often fixing incorrect reasoning.”

The experts used a dataset of 1,000 curated questions and answers to train its model in less than 30 minutes using Nvidia H100 GPUs, demonstrating that it’s possible to achieve advanced results with a small database and taking advantage of other technologies and AI models.

“Recent advances in reasoning, such as OpenAi’s o1 and DeepSeek’s R1, lack transparency, limiting broader research progress,” wrote the researchers. “Our work aims to push the frontier of reasoning in a fully open manner, fostering innovation and collaboration to accelerate advancements that ultimately benefit society.”

ByteDance Introduces OmniHuman-1, One Of The Most Realistic DeepFake Tool In The Market - 2

Photo by Ben Collins on Unsplash

ByteDance Introduces OmniHuman-1, One Of The Most Realistic DeepFake Tool In The Market

  • Written by Andrea Miliani Former Tech News Expert
  • Fact-Checked by Justyn Newman Former Lead Cybersecurity Editor

Researchers from ByteDance, Tiktok’s parent company, introduced this week a new AI tool called OmniHuman to generate human videos based on images and other media.

In a Rush? Here are the Quick Facts!

  • ByteDance introduced its latest AI tool, OmniHuman-1, which can generate videos with realistic motion, style, and behavior from a single photo.
  • The research team shared a paper with more details of the methodology and strategies applied to reach realistic deepfakes.
  • OmniHuman is not available to the public yet.

The first version of the AI tool, OmniHuman-1, is capable of generating videos that support multiple image styles—ranging from realistic photography to animation and cartoons—as well as offering audio and music variations, various aspect ratios, and realistic motion images. In several demos, the Chinese company showcased the AI model’s capabilities.

“OmniHuman significantly outperforms existing methods, generating extremely realistic human videos based on weak signal inputs, especially audio,” states the paper published on Monday by Bytedance’s team. “ It supports image inputs of any aspect ratio, whether they are portraits, half-body, or full-body images, delivering more lifelike and high-quality results across various scenarios.”

The research team explained that they used a “multimodality motion conditioning mixed training strategy” and provided several examples of the tool’s capabilities, including recreating a class with Albert Einstein, simulating speeches using images from royalty-free websites, and generating musical performances from audio or video media.

ByteDance’s research team warned about the risks of fraud—they haven’t released the AI tool to the public yet, and didn’t share a date—and other ethics concerns. The company assured that the images and videos used to demonstrate the model’s performance were taken from public sources.

According to Forbes , the Chinese company used 18,700 hours of human video data to train the new mode. Multiple experts have already shared their thoughts on the new AI tool.

“Creating something from just a picture and making it look like it’s really talking and really moving is fascinating from a technological standpoint, but it could have a lot of potential negative consequences, too,” said Samantha G. Wolfe, an adjunct professor at NYU’s Steinhardt School of Culture, Education and Human Development and founder of PitchFWD in an interview with Forbes. “Pretend versions of business leaders or political leaders saying something that isn’t accurate can have a huge influence on a business, or a huge influence on a country.”

Wolfe’s concerns are shared by multiple experts in the industry. Cybersecurity experts recently warned about a new wave of scams with sophisticated AI-generated deepfakes.