{"id":833357,"date":"2025-03-12T04:11:50","date_gmt":"2025-03-12T09:11:50","guid":{"rendered":"https:\/\/newsycanuse.com\/index.php\/2025\/03\/12\/google-calls-gemma-3-the-most-powerful-ai-model-you-can-run-on-one-gpu\/"},"modified":"2025-03-12T04:11:50","modified_gmt":"2025-03-12T09:11:50","slug":"google-calls-gemma-3-the-most-powerful-ai-model-you-can-run-on-one-gpu","status":"publish","type":"post","link":"https:\/\/newsycanuse.com\/index.php\/2025\/03\/12\/google-calls-gemma-3-the-most-powerful-ai-model-you-can-run-on-one-gpu\/","title":{"rendered":"Google calls Gemma 3 the most powerful AI model you can run on one GPU"},"content":{"rendered":"<div>\n<div>\n<p><img alt=\"Richard Lawler\" data-chromatic=\"ignore\" loading=\"lazy\" width=\"36\" height=\"36\" decoding=\"async\" data-nimg=\"1\"  src=\"https:\/\/platform.theverge.com\/wp-content\/uploads\/sites\/2\/2025\/01\/0007835275_20.jpg?quality=90&#038;strip=all&#038;crop=0%2C0%2C100%2C100&#038;w=96\"><\/p>\n<p><a href=\"http:\/\/www.theverge.com\/authors\/richard-lawler\">Richard Lawler<\/a> <span>is a senior editor following news across tech, culture, policy, and entertainment. He joined The Verge in 2021 after several years covering news at Engadget.<\/span><\/p>\n<\/div>\n<div id=\"zephr-anchor\">\n<p>A little over a year after <a href=\"http:\/\/www.theverge.com\/2024\/2\/21\/24078610\/google-gemma-gemini-small-ai-model-open-source\">releasing two \u201copen\u201d Gemma AI models<\/a> built from the same technology behind its Gemini AI, Google is <a href=\"http:\/\/blog.google\/technology\/developers\/gemma-3\">updating the family with Gemma 3<\/a>. According to the blog post, these models are intended for use by developers creating AI applications capable of running wherever they\u2019re needed, on anything from a phone to a workstation with support for over 35 languages, as well as the ability to analyze text, images, and short videos.<\/p>\n<p>The company claims that it\u2019s the \u201cworld\u2019s best single-accelerator model,\u201d outperforming competition from Facebook\u2019s Llama, <a href=\"http:\/\/www.theverge.com\/24353060\/deepseek-ai-china-nvidia-openai\">DeepSeek<\/a>, and OpenAI for performance on a host with a single GPU, as well as optimized capabilities for running on Nvidia\u2019s GPUs and dedicated AI hardware. Gemma 3\u2019s vision encoder is also upgraded, with support for high-res and non-square images, while the new <a href=\"https:\/\/developers.googleblog.com\/en\/safer-and-multimodal-responsible-ai-with-gemma\/\">ShieldGemma 2 image safety classifier<\/a> is available for use to filter both image input and output for content classified as sexually explicit, dangerous, or violent.<\/p>\n<div>\n<p>To go deeper into those claims, you can check out the <a href=\"https:\/\/storage.googleapis.com\/deepmind-media\/gemma\/Gemma3Report.pdf\">26-page technical report<\/a>.<\/p>\n<\/div>\n<div>\n<div class>\n<p><a href=\"https:\/\/platform.theverge.com\/wp-content\/uploads\/sites\/2\/2025\/03\/Chatbot_Arena_ELO_Score.width-1000.format-webp.webp?quality=90&#038;strip=all&#038;crop=2.375,0,95.25,100\" data-pswp-height=\"635\" data-pswp-width=\"952.4999999999999\" target=\"_blank\" rel=\"noreferrer\"><img alt=\"Graph showing Chatbot arena elo scores ranked, with gemma coming in second while requiring only one Nvidia H100 GPU.\" data-chromatic=\"ignore\" loading=\"lazy\" decoding=\"async\" data-nimg=\"fill\"   src=\"https:\/\/platform.theverge.com\/wp-content\/uploads\/sites\/2\/2025\/03\/Chatbot_Arena_ELO_Score.width-1000.format-webp.webp?quality=90&#038;strip=all&#038;crop=2.375%2C0%2C95.25%2C100&#038;w=2400\"><\/a><\/p>\n<\/div>\n<p><cite>Image: Google<\/cite><\/p>\n<\/div>\n<p>Last year it was unclear how much interest there would be in a model like Gemma, however, <a href=\"http:\/\/www.theverge.com\/ai-artificial-intelligence\/598846\/deepseek-big-tech-ai-industry-nvidia-impac\">the popularity of DeepSeek and others<\/a> shows there is interest in AI tech with lower hardware requirements. Despite its claims of advanced capabilities, Google also says, \u201cGemma 3\u2019s enhanced STEM performance prompted specific evaluations focused on its potential for misuse in creating harmful substances; their results indicate a low risk level.\u201d<\/p>\n<p>What exactly constitutes an \u201copen\u201d or \u201copen source\u201d AI model remains a topic of debate, and with Google\u2019s Gemma, that has focused on the company\u2019s license that restricts what people are allowed to use it for, which has not changed with this new release. Google is continuing to promote Gemma with Google Cloud credits, and the Gemma 3 Academic program will allow academic researchers to apply for $10,000 worth of credits to accelerate their research.<\/p>\n<\/div>\n<\/div>\n<p><a href=\"https:\/\/www.theverge.com\/ai-artificial-intelligence\/627968\/google-gemma-3-open-ai-model\" class=\"button purchase\" rel=\"nofollow noopener\" target=\"_blank\">Read More<\/a><br \/>\n Richard Lawler<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Richard Lawler is a senior editor following news across tech, culture, policy, and entertainment. He joined The Verge in 2021 after several years covering news at Engadget. A little over a year after releasing two \u201copen\u201d Gemma AI models built from the same technology behind its Gemini AI, Google is updating the family with Gemma<\/p>\n","protected":false},"author":1,"featured_media":833358,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[127,496,46],"tags":[],"class_list":{"0":"post-833357","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-calls","8":"category-google","9":"category-technology"},"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/posts\/833357","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/comments?post=833357"}],"version-history":[{"count":0,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/posts\/833357\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/media\/833358"}],"wp:attachment":[{"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/media?parent=833357"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/categories?post=833357"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/tags?post=833357"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}