The cerita lucah melayu terkini kongkek janda gersangfloodgates 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.
EWP Announces New Artistic Director
Founder of embattled Chinese EV maker reportedly moves to the US · TechNode
Best TV deal: Save $40 on Hisense 75
How to unblock Xnxx for free
Gate.io Implements Zero
Google Search makes AI Mode available for everyone to try out
EV maker WM Motor suspends in
Google Search makes AI Mode available for everyone to try out
The Second Edition of ETH Belgrade Emerges This June
EV maker WM Motor suspends in
'Mele Murals' to Premiere Friday at Aratani
HoYoverse fails in attempts to bypass App Store’s 30% "Apple Tax" · TechNode
GAC’s newest falcon
The Terrorist Next Door
Pasadena Chalk Festival This Weekend
Sabalenka vs. Svitolina 2025 livestream: Watch Madrid Open for free
NYT Strands hints, answers for May 1
US tightens export restrictions on AI chips to China · TechNode
‘PV Stories’ Reception and Screening
Vivo set to unveil self
Everything coming to Amazon Prime Video in JulyThe 12 best tweets of the week, including Beethoven, dry ribs, and BatmanNuro delivers FedEx packages in driverless vehiclesHBO Max sends out mysterious email, everyone makes the same jokeHubble snaps a radiant galaxy lit up by a very active black holeGoogle Assistant adds Black history feature in honor of JuneteenthPixar's 'Luca' is the ultimate summer vacation fantasy: Movie reviewEverything coming to Netflix in JulyHow to save an iPhone live photo as a video10 years on, 'Kerbal Space Program' shoots for more distant stars The bitter, banal, and bizarre YouTube circus of Depp v. Heard What is digital minimalism and how can it help you reclaim your time 5 Days of Awesome Wallpapers: Geometric and Architectural Wallpapers Roe v. Wade overturned by Supreme Court in historic decision 'Sex Bomb' celebrates the joy of embracing sexuality and love as a British Hanger challenge on TikTok tests odd head reflex Damon Dominique on teaching French and the importance of saying 'Um' The 11 best and funniest tweets of the week, including crab rangoon and gargoyles Technoblade, 'Minecraft' YouTube creator, dies aged 23, family and fans pay tribute The 10 best and funniest tweets of the week, including iPods and Nick Lachey
0.1893s , 10090.21875 kb
Copyright © 2025 Powered by 【cerita lucah melayu terkini kongkek janda gersang】Researchers created an AI reasoning model on par with OpenAI's o1 for less than $50,Feature Flash
