{"id":181906,"date":"2026-06-05T14:48:14","date_gmt":"2026-06-05T12:48:15","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/amd-gaia-0-20-0"},"modified":"2026-06-05T14:48:14","modified_gmt":"2026-06-05T12:48:15","slug":"amd-gaia-0-20-0","status":"publish","type":"post","link":"https:\/\/prohoster.info\/pt\/blog\/news\/amd-gaia-0-20-0","title":{"rendered":"AMD GAIA 0.20.0","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"<p>Lan\u00e7amento publicado <strong>AMD GAIA 0.20.0<\/strong> \u2014 framework aberto para executar agentes de IA locais em PCs com acelera\u00e7\u00e3o de hardware AMD Ryzen AI. O projeto \u00e9 distribu\u00eddo sob licen\u00e7a <strong>MIT<\/strong>, suportando Windows e Linux, e a instala\u00e7\u00e3o est\u00e1 dispon\u00edvel atrav\u00e9s do pacote amd-gaia. A tag v0.20.0 foi publicada em 3 de junho, mas chegou ao feed de not\u00edcias em 4\u20135 de junho.<\/p>\n<p>A principal mudan\u00e7a da vers\u00e3o \u00e9 a escolha adequada do dispositivo de execu\u00e7\u00e3o para cada agente. Anteriormente, o GAIA usava por padr\u00e3o GPU atrav\u00e9s de um backend baseado em llama.cpp e n\u00e3o oferecia uma maneira conveniente de alternar um agente espec\u00edfico para CPU ou NPU de Ryzen, que \u00e9 energeticamente eficiente. No GAIA 0.20.0, os agentes podem declarar os dispositivos suportados, e o usu\u00e1rio pode escolher entre CPU, GPU ou NPU atrav\u00e9s da interface do agente (Agent UI) ou pela op\u00e7\u00e3o de CLI \u2014device {cpu,gpu,npu}. A GPU permanece como a op\u00e7\u00e3o padr\u00e3o, enquanto o perfil gaia init \u2014profile npu se encarrega da detec\u00e7\u00e3o do NPU, instala\u00e7\u00e3o do FLM-backend e carregamento do modelo.<\/p>\n<p>Mudan\u00e7as na vers\u00e3o:<\/p>\n<ul>\n<li>\n<p><strong>Sele\u00e7\u00e3o de CPU\/GPU\/NPU para agentes individuais.<\/strong> Permite executar cen\u00e1rios pesados na GPU e cen\u00e1rios menos exigentes ou em segundo plano na NPU ou CPU. Para os propriet\u00e1rios de Ryzen AI, isso \u00e9 especialmente importante: o NPU pode ser usado para sa\u00edda local com menor consumo de energia, sem ocupar a unidade de processamento gr\u00e1fico.<\/p>\n<\/li>\n<li>\n<p><strong>Agent Hub TUI.<\/strong> Ao executar o gaia sem argumentos, agora abre o centro de controle de agentes em terminal. Atrav\u00e9s dele, \u00e9 poss\u00edvel visualizar, buscar, iniciar e administrar agentes sem uma interface gr\u00e1fica. Os desenvolvedores indicam que o tamanho do bin\u00e1rio aut\u00f4nomo \u00e9 de cerca de 21 MB e o tempo de inicializa\u00e7\u00e3o \u00e9 de menos de 200 ms.<\/p>\n<\/li>\n<li>\n<p><strong>Controle mais rigoroso das ferramentas MCP.<\/strong> Para conectores MCP, foi adicionado um segundo n\u00edvel de controle \u2014 activations. Agora, o conector atribu\u00eddo a um agente n\u00e3o precisa mais adicionar automaticamente todas as suas ferramentas no prompt: as ferramentas se tornam vis\u00edveis apenas ap\u00f3s serem explicitamente ativadas para o par \"conector - agente\". Isso reduz o ru\u00eddo no prompt e ajuda pequenos modelos a escolherem a\u00e7\u00f5es espec\u00edficas de forma mais precisa.<\/p>\n<\/li>\n<li>\n<p><strong>Processamento de e-mails acelerado.<\/strong> O agente de e-mail recebeu sete novas ferramentas de batch para a organiza\u00e7\u00e3o em massa dos itens recebidos. De acordo com os desenvolvedores, o cen\u00e1rio t\u00edpico foi reduzido de cerca de 13 solicita\u00e7\u00f5es de LLM para 2\u20133 etapas, o tempo de processamento foi de 488 segundos para 30\u201360 segundos, e o consumo de tokens foi reduzido de aproximadamente 12 mil para 1,2 mil.<\/p>\n<\/li>\n<li>\n<p><strong>RAG para arquivos PowerPoint.<\/strong> A GAIA agora indexa diretamente .pptx: texto, tabelas, notas do orador e imagens embutidas atrav\u00e9s da an\u00e1lise VLM. Antes, o usu\u00e1rio precisava primeiro salvar as apresenta\u00e7\u00f5es em PDF.<\/p>\n<\/li>\n<li>\n<p><strong>Refor\u00e7o da seguran\u00e7a e resili\u00eancia do primeiro lan\u00e7amento.<\/strong> A vers\u00e3o resolve o contorno da prote\u00e7\u00e3o contra grava\u00e7\u00e3o atrav\u00e9s de links simb\u00f3licos no Python 3.10\/3.11, estendendo as restri\u00e7\u00f5es de grava\u00e7\u00e3o para mais quatro ferramentas de arquivo, al\u00e9m de corrigir um erro de diagn\u00f3stico de modelo corrompido que poderia levar \u00e0 relut\u00e2ncia de cerca de 25 GB de dados.<\/p>\n<\/li>\n<\/ul>\n<p>A GAIA se desenvolve como uma alternativa local aos servi\u00e7os de IA em nuvem: os dados permanecem na m\u00e1quina do usu\u00e1rio, e a execu\u00e7\u00e3o pode ser transferida entre CPU, GPU e NPU, dependendo da tarefa. Para a AMD, isso tamb\u00e9m demonstra a aplica\u00e7\u00e3o pr\u00e1tica do Ryzen AI n\u00e3o apenas como um bloco de marketing no processador, mas tamb\u00e9m como um dispositivo computacional separado para agentes locais.<\/p>\n<p>Fonte: <a \ncontent=\"nofollow\" rel=\"nofollow\" href=\"https:\/\/www.linux.org.ru\/news\/ai\/18311463\">linux.org.ru<\/a><\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"<p>\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u0432\u044b\u043f\u0443\u0441\u043a AMD GAIA 0.20.0 \u2014 \u043e\u0442\u043a\u0440\u044b\u0442\u043e\u0433\u043e \u0444\u0440\u0435\u0439\u043c\u0432\u043e\u0440\u043a\u0430 \u0434\u043b\u044f \u0437\u0430\u043f\u0443\u0441\u043a\u0430 \u043b\u043e\u043a\u0430\u043b\u044c\u043d\u044b\u0445 AI-\u0430\u0433\u0435\u043d\u0442\u043e\u0432 \u043d\u0430 \u041f\u041a \u0441 \u0430\u043f\u043f\u0430\u0440\u0430\u0442\u043d\u044b\u043c \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u0438\u0435\u043c AMD Ryzen AI. \u041f\u0440\u043e\u0435\u043a\u0442 \u0440\u0430\u0441\u043f\u0440\u043e\u0441\u0442\u0440\u0430\u043d\u044f\u0435\u0442\u0441\u044f \u043f\u043e\u0434 \u043b\u0438\u0446\u0435\u043d\u0437\u0438\u0435\u0439 MIT, \u043f\u043e\u0434\u0434\u0435\u0440\u0436\u0438\u0432\u0430\u0435\u0442 Windows \u0438 Linux, \u0430 \u0443\u0441\u0442\u0430\u043d\u043e\u0432\u043a\u0430 \u0434\u043e\u0441\u0442\u0443\u043f\u043d\u0430 \u0447\u0435\u0440\u0435\u0437 \u043f\u0430\u043a\u0435\u0442 amd-gaia. \u0421\u0430\u043c \u0442\u0435\u0433 v0.20.0 \u043e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d 3 \u0438\u044e\u043d\u044f, \u043d\u043e \u0432 \u043d\u043e\u0432\u043e\u0441\u0442\u043d\u0443\u044e \u043b\u0435\u043d\u0442\u0443 \u0440\u0435\u043b\u0438\u0437 \u043f\u043e\u043f\u0430\u043b 4\u20135 \u0438\u044e\u043d\u044f. \u0413\u043b\u0430\u0432\u043d\u043e\u0435 \u0438\u0437\u043c\u0435\u043d\u0435\u043d\u0438\u0435 \u0432\u0435\u0440\u0441\u0438\u0438 \u2014 \u043d\u043e\u0440\u043c\u0430\u043b\u044c\u043d\u044b\u0439 \u0432\u044b\u0431\u043e\u0440 [&hellip;]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":8,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[702],"tags":[],"class_list":["post-181906","post","type-post","status-publish","format-standard","hentry","category-news"],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.3 - aioseo.com -->\n\t<meta name=\"description\" content=\"\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u0432\u044b\u043f\u0443\u0441\u043a AMD GAIA 0.20.0 \u2014 \u043e\u0442\u043a\u0440\u044b\u0442\u043e\u0433\u043e \u0444\u0440\u0435\u0439\u043c\u0432\u043e\u0440\u043a\u0430 \u0434\u043b\u044f \u0437\u0430\u043f\u0443\u0441\u043a\u0430 \u043b\u043e\u043a\u0430\u043b\u044c\u043d\u044b\u0445 AI-\u0430\u0433\u0435\u043d\u0442\u043e\u0432 \u043d\u0430 \u041f\u041a \u0441 \u0430\u043f\u043f\u0430\u0440\u0430\u0442\u043d\u044b\u043c \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u0438\u0435\u043c AMD Ryzen AI.\" \/>\n\t<meta name=\"robots\" content=\"max-image-preview:large\" \/>\n\t<meta name=\"author\" content=\"Erik Peterson\"\/>\n\t<link rel=\"canonical\" href=\"https:\/\/prohoster.info\/pt\/blog\/news\/amd-gaia-0-20-0\" \/>\n\t\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.3\" \/>\n\t\t<meta property=\"og:locale\" content=\"pt_PT\" \/>\n\t\t<meta property=\"og:site_name\" content=\"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b\" \/>\n\t\t<meta property=\"og:type\" content=\"article\" \/>\n\t\t<meta property=\"og:title\" content=\"\ud83e\udd47AMD GAIA 0.20.0 | ProHoster\" \/>\n\t\t<meta property=\"og:description\" content=\"\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u0432\u044b\u043f\u0443\u0441\u043a AMD GAIA 0.20.0 \u2014 \u043e\u0442\u043a\u0440\u044b\u0442\u043e\u0433\u043e \u0444\u0440\u0435\u0439\u043c\u0432\u043e\u0440\u043a\u0430 \u0434\u043b\u044f \u0437\u0430\u043f\u0443\u0441\u043a\u0430 \u043b\u043e\u043a\u0430\u043b\u044c\u043d\u044b\u0445 AI-\u0430\u0433\u0435\u043d\u0442\u043e\u0432 \u043d\u0430 \u041f\u041a \u0441 \u0430\u043f\u043f\u0430\u0440\u0430\u0442\u043d\u044b\u043c \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u0438\u0435\u043c AMD Ryzen AI.\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/prohoster.info\/pt\/blog\/news\/amd-gaia-0-20-0\" \/>\n\t\t<meta property=\"og:image\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:secure_url\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:width\" content=\"350\" \/>\n\t\t<meta property=\"og:image:height\" content=\"350\" \/>\n\t\t<meta property=\"article:published_time\" content=\"2026-06-05T12:48:15+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2026-06-05T12:48:15+00:00\" \/>\n\t\t<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<meta property=\"article:author\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<!-- All in One SEO -->\n\n","aioseo_head_json":{"title":"\ud83e\udd47AMD GAIA 0.20.0 | ProHoster","description":"Foi lan\u00e7ado o AMD GAIA 0.20.0 \u2014 uma estrutura de c\u00f3digo aberto para executar agentes de IA locais em PCs com acelera\u00e7\u00e3o de hardware AMD Ryzen AI.","canonical_url":"https:\/\/prohoster.info\/pt\/blog\/news\/amd-gaia-0-20-0","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"pt_PT","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd47AMD GAIA 0.20.0 | ProHoster","og:description":"\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u0432\u044b\u043f\u0443\u0441\u043a AMD GAIA 0.20.0 \u2014 \u043e\u0442\u043a\u0440\u044b\u0442\u043e\u0433\u043e \u0444\u0440\u0435\u0439\u043c\u0432\u043e\u0440\u043a\u0430 \u0434\u043b\u044f \u0437\u0430\u043f\u0443\u0441\u043a\u0430 \u043b\u043e\u043a\u0430\u043b\u044c\u043d\u044b\u0445 AI-\u0430\u0433\u0435\u043d\u0442\u043e\u0432 \u043d\u0430 \u041f\u041a \u0441 \u0430\u043f\u043f\u0430\u0440\u0430\u0442\u043d\u044b\u043c \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u0438\u0435\u043c AMD Ryzen AI.","og:url":"https:\/\/prohoster.info\/pt\/blog\/news\/amd-gaia-0-20-0","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2026-06-05T12:48:15+00:00","article:modified_time":"2026-06-05T12:48:15+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":[],"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/pt\/wp-json\/wp\/v2\/posts\/181906","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/pt\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/pt\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/pt\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/pt\/wp-json\/wp\/v2\/comments?post=181906"}],"version-history":[{"count":0,"href":"https:\/\/prohoster.info\/pt\/wp-json\/wp\/v2\/posts\/181906\/revisions"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/pt\/wp-json\/wp\/v2\/media?parent=181906"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/pt\/wp-json\/wp\/v2\/categories?post=181906"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/pt\/wp-json\/wp\/v2\/tags?post=181906"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}