The sasha foxxx brain rewired to eroticize debtfloodgates have opened for building AI reasoning models on the cheap.
Researchers at Stanford and the University of Washington have developed a model that performs comparably to OpenAI o1 and DeepSeek R1 models in math and coding — for less than $50 of cloud compute credits.
What's more, the model was trained on only 1,000 questions, and took just 26 minutes and 16 Nvidia H100 GPUs. Stanford researcher Niklas Muennighoff said in a email to Mashable that the cost is an estimate based on the GPU runtime and number of H100 GPUs used.
The AI industry of late is all about how new approaches to the pre and post training process can massively save computing costs, as evidenced by DeepSeek's disruptive impact. On top of that, developers are now able to build on top of existing AI models at little or no cost, through APIs, open-source access, and even closed-source models by distilling their data, bringing the costs down even more.
According to the team's research paper which was published last Friday, s1 was trained on a dataset consisting of "1,000 carefully curated questions paired with reasoning traces and answers distilled from Gemini Thinking Experimental." Google's Gemini Thinking Experimental model is accessible with daily limits through AI Studio. While it's a closed-source model, that clearly hasn't stopped researchers from making use of its responses.
SEE ALSO: OpenAI launches 'deep research' AI agent for ChatGPTNext, the researchers used an "off the shelf" pretrained model from Alibaba-owned lab, Qwen, and performed supervised fine-tuning of its curated dataset. Then, the team created a token budget to control the amount of compute time for testing the model. If s1 went over budget on thinking tokens, it was cut off and forced to generate whatever answer it came up with. If the researchers wanted the model to spend more "test-time compute" on a problem, they would simply tell the model to "wait," which extended its thinking time and led to more accurate results.
By controlling the amount of time and compute spent on a problem, the researchers were able to show how increased thinking team leads to improved performance.
S1 is one example of open-source reasoning models that have been developed for a fraction of the cost of flagship models from Google and OpenAI. In January, UC Berkeley researchers released an open-source reasoning model called Sky-T1 that cost $450, "demonstrating that it is possible to replicate high-level reasoning capabilities affordably and efficiently," per its blog post. There's also the open-source rStar-Math reasoning model from Microsoft Asia researchers, Tulu 3 from non profit research institute Ai2, and HuggingFace has its own initiative to replicate DeepSeek's R1.
As high-quality models become more accessible and cheaper, we're starting to see a power shift from the few AI heavy hitters, to the many.
Topics Artificial Intelligence OpenAI
Three Letters for beyond the Walls by Caio Fernando AbreuWild Apples by Lauren GroffThe Review’s Review: A Happy Pig by The Paris ReviewEverything announced at Samsung Unpacked, including Galaxy AI and Galaxy RingWhat is TikTok's 'orange peel theory'?Redux: Collapse Distinctions by The Paris ReviewSamsung Galaxy S24 Ultra handsOn the Alert for Omens: Rereading Charles Portis by Rosa LysterNew, Tender, Quick: A Visit to the Elizabeth Bishop House by Henri ColeRedux: No Human Tongue by The Paris ReviewJim Jarmusch’s Collages by Lucy SanteAll You Have to Do Is Die by Rowan Hisayo BuchananNYT's The Mini crossword answers for January 18All You Have to Do Is Die by Rowan Hisayo BuchananHunter’s Moon by Nina MacLaughlinMoral Suasion by The Paris ReviewHarvest Moon by Nina MacLaughlinThe Paris Review Podcast Returns by The Paris ReviewSamsung Galaxy S24 Ultra handsMy Father’s Mariannes by Aisha Sabatini Sloan 100 Years Ago, Cinema Saw Its First Nude In the bathroom at a party edits started as a meme. Now they're a sub Prank Idea: Abbots Bromley Horn Dance Looking back on 'Lake Mungo,' must Best TV deal: Amazon Fire TV 50 Tonight: Rowan Ricardo Phillips at McNally Jackson Steve Gianakos: Chubby Boys and Chubby Girls Why is everyone on TikTok doing math problems? 'The Royal Hotel' review: An intense feminist road trip that takes one wrong turn Changing my Slack sound to 'Hummus' made me less stressed Doormat, or, A Story of Charity Season TikTok's brownie recipe comment, explained 'The Exorcist: Believer' review: This legacy sequel is so dull it's a sin Listen to a William Carlos Williams Radio Interview from 1950 “The Dog Wants His Dinner,” a Poem by James Schuyler Why the the New York Times crossword jingle fills us with so much joy It’s Carving Time: Thanksgiving Advice from the 1950s 'Bring Brittney Griner home,' Ben Proudfoot tells President Biden at Oscars 'Bridgerton' Season 2 is the most talked I Thought My Dad Had No More Secrets to Tell, But...
2.8455s , 8288.3984375 kb
Copyright © 2025 Powered by 【sasha foxxx brain rewired to eroticize debt】,Co-creation Information Network