{"id":16025,"date":"2024-07-21T08:50:43","date_gmt":"2024-07-21T00:50:43","guid":{"rendered":"https:\/\/www.1ai.net\/?p=16025"},"modified":"2024-07-21T08:50:43","modified_gmt":"2024-07-21T00:50:43","slug":"%e4%b8%80%e6%96%87%e5%b8%a6%e4%bd%a0%e4%ba%86%e8%a7%a3openai%e6%96%b0%e6%8e%a8%e5%87%ba%e7%9a%84gpt-4o-mini%e6%a8%a1%e5%9e%8b","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/16025.html","title":{"rendered":"This article will introduce you to OpenAI\u2019s newly launched GPT-4o mini model"},"content":{"rendered":"<p data-track=\"1\" data-pm-slice=\"0 0 []\">July 18, 2024<a href=\"https:\/\/www.1ai.net\/en\/tag\/openai\" title=\"[View articles tagged with [OpenAI]]\" target=\"_blank\" >OpenAI<\/a>Launches its most cost-effective model<a href=\"https:\/\/www.1ai.net\/en\/tag\/gpt-4o-mini\" title=\"[See article with [GPT-4o Mini] label]\" target=\"_blank\" >GPT-4o mini<\/a>, what is it used for? How is it different from GPT4-o?<\/p>\n<div class=\"pgc-img\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-16026\" title=\"get-589\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/07\/get-589.jpg\" alt=\"get-589\" width=\"1080\" height=\"929\" \/><\/div>\n<p data-track=\"3\">To sum up, there are three important points:<\/p>\n<ul>\n<li data-track=\"4\"><strong>Lower prices;<\/strong><\/li>\n<li data-track=\"5\"><strong>Faster speed;<\/strong><\/li>\n<li data-track=\"6\"><strong>Excellent effect; (<a href=\"https:\/\/www.1ai.net\/en\/tag\/gpt-4o\" title=\"[View articles tagged with [GPT-4o]]\" target=\"_blank\" >GPT-4o<\/a>Only slightly behind, still superior to most other mainstream models)<\/strong><\/li>\n<\/ul>\n<h1 class=\"pgc-h-arrow-right\" spellcheck=\"false\" data-track=\"9\">Lower price<\/h1>\n<p data-track=\"10\">From the official quotes of the model&#039;s input and output tokens, GPT-4o&#039;s million input tokens are US$5.00, while GPT-4o mini is US$0.15, which is only 3% of GPT-4o&#039;s cost. GPT-4o&#039;s million output tokens are US$15.00, while GPT-4o mini is US$0.60, which is only 4% of GPT-4o&#039;s cost.<\/p>\n<p data-track=\"11\">This price reduction is really enough. Compared with the mainstream models of domestic large model manufacturers that have reduced their prices to 1 to 10 yuan for millions of tokens a while ago, it is still very competitive. For details, see the article [AI Dynamics] A picture to understand the timeline of AI manufacturers&#039; large model API price reductions, millions of tokens &quot;rolled&quot; to cabbage prices!<\/p>\n<p data-track=\"12\">The two pictures below are the official quotations of GPT-4o and GPT-4o mini released by OpenAI.<\/p>\n<div class=\"pgc-img\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-16027\" title=\"get-590\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/07\/get-590.jpg\" alt=\"get-590\" width=\"1080\" height=\"679\" \/><\/div>\n<div class=\"pgc-img\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-16028\" title=\"get-591\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/07\/get-591.jpg\" alt=\"get-591\" width=\"1080\" height=\"640\" \/><\/div>\n<p data-track=\"14\">The screenshot below shows the prices of GPT-4o mini and GPT-3.5 Turbo issued by the OpenAI Developer X Platform account over time.<\/p>\n<div class=\"pgc-img\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-16029\" title=\"get-592\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/07\/get-592.jpg\" alt=\"get-592\" width=\"888\" height=\"811\" \/><\/div>\n<h1 class=\"pgc-h-arrow-right\" spellcheck=\"false\" data-track=\"17\">Faster<\/h1>\n<p data-track=\"18\">GPT-4o mini, with its low cost and low latency, supports a wide range of tasks, such as chaining or parallelizing multiple model calls (e.g., calling multiple APIs), passing large amounts of context to models (e.g., a complete code base or conversation history), or interacting with customers through fast, real-time text responses (e.g., a customer support chatbot).<\/p>\n<p data-track=\"20\">However, I have not been able to personally compare and test the speed yet. Before comparing GPT-4o and GPT-4, I intuitively felt that the response speed was indeed much faster.<\/p>\n<h1 class=\"pgc-h-arrow-right\" spellcheck=\"false\" data-track=\"23\">Excellent results<\/h1>\n<p data-track=\"24\">GPT-4o mini outperforms GPT-3.5 Turbo and other small models on academic benchmarks for text intelligence and multimodal reasoning, and supports the same range of languages as GPT-4o.<\/p>\n<p data-track=\"25\">GPT-4o mini has been evaluated on several key benchmarks, which OpenAI describes as follows:<\/p>\n<p data-track=\"26\">Reasoning tasks: GPT-4o mini outperforms other small models on reasoning tasks involving text and vision, scoring 82.0% on the text intelligence and reasoning benchmark MMLU, compared to 77.9% for Gemini Flash and 73.8% for Claude Haiku.<\/p>\n<p data-track=\"27\">Mathematical and Coding Skills: GPT-4o mini performs well on mathematical reasoning and coding tasks, outperforming previous small models on the market. On MGSM, which measures mathematical reasoning, GPT-4o mini scored 87.0%, while Gemini Flash scored 75.5% and Claude Haiku scored 71.7%. On HumanEval, which measures coding performance, GPT-4o mini scored 87.2%, while Gemini Flash scored 71.5% and Claude Haiku scored 75.9%.<\/p>\n<p data-track=\"28\">Multimodal Reasoning: GPT-4o mini also showed strong performance on the multimodal reasoning evaluation MMMU, scoring 59.4%, while Gemini Flash scored 56.1% and Claude Haiku scored 50.2%.<\/p>\n<p data-track=\"29\">Below is a screenshot of the benchmark evaluation results.<\/p>\n<div class=\"pgc-img\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-16030\" title=\"get-593\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/07\/get-593.jpg\" alt=\"get-593\" width=\"1080\" height=\"608\" \/><\/div>\n<p>&nbsp;<\/p>","protected":false},"excerpt":{"rendered":"<p>On July 18, 2024, OpenAI launched its most cost-effective model, the GPT-4o mini, what does it do? How is it different from GPT4-o? Summarized below are the following 3 important points: lower price; faster speed; excellent results; (only a little worse than GPT-4o, and still better than most other mainstream models) Lower price From the model's input and output tokens official quotes, the GPT-4o's millions of input tokens for US$5.00, while the GPT-4o mini for GPT-4o's million input tokens are US$5.00, while GPT-4o mini is US$0.15, which is only 3% of GPT-4o's cost.GPT-4o's million output tokens are US$15.00, while GPT-4o mini is US$0.60, which is only<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[144],"tags":[2582,3578,190],"collection":[],"class_list":["post-16025","post","type-post","status-publish","format-standard","hentry","category-baike","tag-gpt-4o","tag-gpt-4o-mini","tag-openai"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/16025","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=16025"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/16025\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=16025"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=16025"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=16025"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=16025"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}