{"id":949,"date":"2026-03-07T04:30:26","date_gmt":"2026-03-06T19:30:26","guid":{"rendered":"https:\/\/itexplore.org\/jp\/columns\/gpt-5-4-luma-ai-agents-graviton-amplitudes-ai-tech-trends\/"},"modified":"2026-03-07T04:30:26","modified_gmt":"2026-03-06T19:30:26","slug":"gpt-5-4-luma-ai-agents-graviton-amplitudes-ai-tech-trends","status":"publish","type":"post","link":"https:\/\/itexplore.org\/jp\/columns\/gpt-5-4-luma-ai-agents-graviton-amplitudes-ai-tech-trends\/","title":{"rendered":"GPT-5.4\u3001Luma AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3001\u91cd\u529b\u5b50\u632f\u5e45\uff1a\u6700\u65b0AI\u6280\u8853\u52d5\u5411"},"content":{"rendered":"<p>\u672c\u65e5\u306e\u6ce8\u76eeAI\u30fb\u30c6\u30c3\u30af\u30cb\u30e5\u30fc\u30b9\u3092\u3001\u5c02\u9580\u7684\u306a\u5206\u6790\u3068\u5171\u306b\u304a\u5c4a\u3051\u3057\u307e\u3059\u3002<\/p>\n<div class=\"wp-block-vk-blocks-alert vk_alert alert alert-warning has-alert-icon\">\n<div class=\"vk_alert_icon\">\n<div class=\"vk_alert_icon_icon\"><i class=\"fa-solid fa-triangle-exclamation\" aria-hidden=\"true\"><\/i><\/div>\n<div class=\"vk_alert_icon_text\"><span>Warning<\/span><\/div>\n<\/div>\n<div class=\"vk_alert_content\">\n<p>\u3053\u306e\u8a18\u4e8b\u306fAI\u306b\u3088\u3063\u3066\u81ea\u52d5\u751f\u6210\u30fb\u5206\u6790\u3055\u308c\u305f\u3082\u306e\u3067\u3059\u3002AI\u306e\u6027\u8cea\u4e0a\u3001\u4e8b\u5b9f\u8aa4\u8a8d\u304c\u542b\u307e\u308c\u308b\u53ef\u80fd\u6027\u304c\u3042\u308b\u305f\u3081\u3001\u91cd\u8981\u306a\u5224\u65ad\u3092\u4e0b\u3059\u969b\u306f\u5fc5\u305a\u30ea\u30f3\u30af\u5148\u306e\u4e00\u6b21\u30bd\u30fc\u30b9\u3092\u3054\u78ba\u8a8d\u304f\u3060\u3055\u3044\u3002<\/p>\n<\/div>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">\u30d7\u30ed\u30d5\u30a7\u30c3\u30b7\u30e7\u30ca\u30eb\u306a\u4f5c\u696d\u5411\u3051\u306b\u8a2d\u8a08\u3055\u308c\u305fGPT-5.4\u306e\u7d39\u4ecb<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> Introducing GPT-5.4 | OpenAI<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p><strong>OpenAI<\/strong>\u306f\u3001\u30d7\u30ed\u30d5\u30a7\u30c3\u30b7\u30e7\u30ca\u30eb\u306a\u4f5c\u696d\u5411\u3051\u306b\u8a2d\u8a08\u3055\u308c\u305f\u6700\u65b0\u306e\u30d5\u30ed\u30f3\u30c6\u30a3\u30a2\u30e2\u30c7\u30eb<strong>GPT-5.4<\/strong>\u3092\u767a\u8868\u3057\u307e\u3057\u305f\u3002\u3053\u306e\u30e2\u30c7\u30eb\u306f\u3001<strong>ChatGPT<\/strong>\uff08GPT-5.4 Thinking\u3068\u3057\u3066\uff09\u3001API\u3001\u304a\u3088\u3073<strong>Codex<\/strong>\u3067\u5229\u7528\u53ef\u80fd\u3067\u3042\u308a\u3001\u8907\u96d1\u306a\u30bf\u30b9\u30af\u3067\u6700\u5927\u306e\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u6c42\u3081\u308b\u30e6\u30fc\u30b6\u30fc\u5411\u3051\u306b<strong>GPT-5.4 Pro<\/strong>\u3082\u63d0\u4f9b\u3055\u308c\u307e\u3059\u3002<\/p>\n<p><strong>GPT-5.4<\/strong>\u306f\u3001\u63a8\u8ad6\u3001\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u3001\u304a\u3088\u3073\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u306b\u304a\u3051\u308b\u6700\u8fd1\u306e\u9032\u6b69\u3092\u7d71\u5408\u3057\u3001<strong>GPT-5.3-Codex<\/strong>\u306e\u696d\u754c\u3092\u30ea\u30fc\u30c9\u3059\u308b\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u80fd\u529b\u3092\u7d44\u307f\u8fbc\u307f\u3064\u3064\u3001\u30c4\u30fc\u30eb\u3001\u30bd\u30d5\u30c8\u30a6\u30a7\u30a2\u74b0\u5883\u3001\u30b9\u30d7\u30ec\u30c3\u30c9\u30b7\u30fc\u30c8\u3001\u30d7\u30ec\u30bc\u30f3\u30c6\u30fc\u30b7\u30e7\u30f3\u3001\u30c9\u30ad\u30e5\u30e1\u30f3\u30c8\u306a\u3069\u306e\u30d7\u30ed\u30d5\u30a7\u30c3\u30b7\u30e7\u30ca\u30eb\u30bf\u30b9\u30af\u3067\u306e\u30e2\u30c7\u30eb\u306e\u52d5\u4f5c\u3092\u6539\u5584\u3057\u3066\u3044\u307e\u3059\u3002\u3053\u308c\u306b\u3088\u308a\u3001\u3088\u308a\u6b63\u78ba\u3067\u52b9\u7387\u7684\u306a\u4f5c\u696d\u304c\u53ef\u80fd\u306b\u306a\u308a\u307e\u3059\u3002\u7279\u306b\u3001<strong>GPT-5.4 Thinking<\/strong>\u306f\u3001\u601d\u8003\u306e\u4e8b\u524d\u8a08\u753b\u3092\u63d0\u4f9b\u3057\u3001\u6df1\u3044\u30a6\u30a7\u30d6\u8abf\u67fb\u80fd\u529b\u3092\u5411\u4e0a\u3055\u305b\u3001\u3088\u308a\u9ad8\u54c1\u8cea\u306a\u56de\u7b54\u3092\u8fc5\u901f\u306b\u63d0\u4f9b\u3057\u307e\u3059\u3002<\/p>\n<p><strong>GPT-5.4<\/strong>\u306f\u3001\u30cd\u30a4\u30c6\u30a3\u30d6\u306a\u6700\u5148\u7aef\u306e\u30b3\u30f3\u30d4\u30e5\u30fc\u30bf\u30fc\u4f7f\u7528\u80fd\u529b\u3092\u5099\u3048\u305f\u521d\u306e\u6c4e\u7528\u30e2\u30c7\u30eb\u3067\u3042\u308a\u3001\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u9593\u3067\u8907\u96d1\u306a\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u3092\u5b9f\u884c\u3067\u304d\u308b\u3088\u3046\u306b\u3057\u307e\u3059\u3002\u6700\u5927<strong>1M\u30c8\u30fc\u30af\u30f3<\/strong>\u306e\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u3092\u30b5\u30dd\u30fc\u30c8\u3057\u3001\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u9577\u671f\u9593\u306b\u308f\u305f\u308b\u30bf\u30b9\u30af\u3092\u8a08\u753b\u3001\u5b9f\u884c\u3001\u691c\u8a3c\u3059\u308b\u3053\u3068\u3092\u53ef\u80fd\u306b\u3057\u307e\u3059\u3002\u307e\u305f\u3001<strong>\u30c4\u30fc\u30eb\u691c\u7d22<\/strong>\u3092\u5c0e\u5165\u3057\u3001\u5927\u898f\u6a21\u306a\u30c4\u30fc\u30eb\u30a8\u30b3\u30b7\u30b9\u30c6\u30e0\u3067\u306e\u52b9\u7387\u7684\u306a\u30c4\u30fc\u30eb\u5229\u7528\u3092\u5b9f\u73fe\u3057\u3001\u30c8\u30fc\u30af\u30f3\u4f7f\u7528\u91cf\u3092\u5927\u5e45\u306b\u524a\u6e1b\u3057\u307e\u3059\u3002<\/p>\n<p>\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u80fd\u529b\u3082\u5f37\u5316\u3055\u308c\u3001<strong>SWE-Bench Pro<\/strong>\u3067<strong>GPT-5.3-Codex<\/strong>\u3068\u540c\u7b49\u4ee5\u4e0a\u306e\u6027\u80fd\u3092\u767a\u63ee\u3057\u3001\u30d5\u30ed\u30f3\u30c8\u30a8\u30f3\u30c9\u30bf\u30b9\u30af\u306b\u304a\u3044\u3066\u3088\u308a\u7f8e\u7684\u3067\u6a5f\u80fd\u7684\u306a\u7d50\u679c\u3092\u751f\u307f\u51fa\u3057\u307e\u3059\u3002\u3055\u3089\u306b\u3001<strong>Playwright (Interactive)<\/strong>\u3068\u3044\u3046\u5b9f\u9a13\u7684\u306a<strong>Codex<\/strong>\u30b9\u30ad\u30eb\u3092\u901a\u3058\u3066\u3001\u30a6\u30a7\u30d6\u304a\u3088\u3073Electron\u30a2\u30d7\u30ea\u306e\u8996\u899a\u7684\u306a\u30c7\u30d0\u30c3\u30b0\u3084\u30c6\u30b9\u30c8\u304c\u53ef\u80fd\u306b\u306a\u308a\u307e\u3059\u3002\u5b89\u5168\u6027\u306b\u95a2\u3057\u3066\u3082\u3001<strong>Preparedness Framework<\/strong>\u306b\u57fa\u3065\u304d\u3001\u30b5\u30a4\u30d0\u30fc\u5b89\u5168\u30b9\u30bf\u30c3\u30af\u3092\u62e1\u5f35\u3057\u3001<strong>Chain-of-Thought (CoT)<\/strong>\u306e\u76e3\u8996\u53ef\u80fd\u6027\u3092\u5411\u4e0a\u3055\u305b\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/openai.com\/index\/introducing-gpt-5-4\/\" target=\"_blank\" rel=\"noopener\">OpenAI \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> GPT-5.4 significantly advances AI capabilities in professional knowledge work, coding, and autonomous agentic workflows with enhanced computer-use, extended context, and improved safety, making it a powerful tool for complex tasks and developer efficiency.<\/li>\n<li><strong>\u8457\u8005:<\/strong> meetpateltech<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p><strong>OpenAI<\/strong> has unveiled <strong>GPT-5.4<\/strong>, its latest frontier model designed for professional work, available in <strong>ChatGPT<\/strong> (as GPT-5.4 Thinking), the API, and <strong>Codex<\/strong>. A more powerful version, <strong>GPT-5.4 Pro<\/strong>, is also offered for users requiring maximum performance on complex tasks.<\/p>\n<p><strong>GPT-5.4<\/strong> integrates recent advancements in reasoning, coding, and agentic workflows, incorporating the industry-leading coding capabilities of <strong>GPT-5.3-Codex<\/strong> while enhancing its performance across tools, software environments, spreadsheets, presentations, and documents. This results in more accurate and efficient task completion. Specifically, <strong>GPT-5.4 Thinking<\/strong> provides upfront thinking plans and improves deep web research, delivering higher-quality answers faster.<\/p>\n<p>This model is the first general-purpose model with native, state-of-the-art computer-use capabilities, enabling agents to execute complex workflows across applications. It supports up to <strong>1M tokens<\/strong> of context, allowing agents to plan, execute, and verify long-horizon tasks. The introduction of <strong>tool search<\/strong> significantly improves efficiency in large tool ecosystems by reducing token usage.<\/p>\n<p><strong>GPT-5.4<\/strong> also boasts enhanced coding abilities, matching or outperforming <strong>GPT-5.3-Codex<\/strong> on <strong>SWE-Bench Pro<\/strong> and producing more aesthetic and functional results for frontend tasks. An experimental <strong>Codex<\/strong> skill, <strong>Playwright (Interactive)<\/strong>, enables visual debugging and testing of web and Electron apps. Safety measures include an expanded cyber safety stack under the <strong>Preparedness Framework<\/strong> and improved <strong>Chain-of-Thought (CoT)<\/strong> monitorability.<\/p>\n<\/blockquote>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">\u5358\u4e00\u30de\u30a4\u30ca\u30b9\u632f\u5e45\u306e\u91cd\u529b\u5b50\u3078\u306e\u62e1\u5f35<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> Extending single-minus amplitudes to gravitons | OpenAI<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p><strong>OpenAI<\/strong>\u306e\u7814\u7a76\u8005\u305f\u3061\u306f\u3001\u91cf\u5b50\u91cd\u529b\u306b\u304a\u3051\u308b\u6563\u4e71\u632f\u5e45\u306b\u95a2\u3059\u308b\u65b0\u3057\u3044\u7814\u7a76\u3092\u767a\u8868\u3057\u307e\u3057\u305f\u3002\u3053\u306e\u7814\u7a76\u306f\u3001\u4ee5\u524d\u30b0\u30eb\u30fc\u30aa\u30f3\u3067\u5f97\u3089\u308c\u305f\u7d50\u679c\u3092\u91cd\u529b\u5b50\u306e\u8a2d\u5b9a\u306b\u62e1\u5f35\u3059\u308b\u3082\u306e\u3067\u3001\u9577\u3089\u304f\u6d88\u6ec5\u3059\u308b\u3068\u8003\u3048\u3089\u308c\u3066\u3044\u305f\u3042\u308b\u7a2e\u306e\u91cd\u529b\u5b50\u76f8\u4e92\u4f5c\u7528\u304c\u3001\u7279\u5b9a\u306e\u904b\u52d5\u5b66\u7684\u6761\u4ef6\u4e0b\u3067\u767a\u751f\u3057\u3046\u308b\u3053\u3068\u3092\u793a\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u3053\u306e\u8ad6\u6587\u300cSingle-minus graviton tree amplitudes are nonzero\u300d\u306f\u3001<strong>Alfredo Guevara<\/strong>\u3001<strong>Alexandru Lupsasca<\/strong>\u3001<strong>David Skinner<\/strong>\u3001<strong>Andrew Strominger<\/strong>\u3001\u304a\u3088\u3073<strong>Kevin Weil<\/strong>\u306b\u3088\u3063\u3066\u57f7\u7b46\u3055\u308c\u307e\u3057\u305f\u3002\u5f7c\u3089\u306f\u3001\u5358\u4e00\u30de\u30a4\u30ca\u30b9\u632f\u5e45\uff081\u3064\u306e\u7c92\u5b50\u304c\u8ca0\u306e\u30d8\u30ea\u30b7\u30c6\u30a3\u3092\u6301\u3061\u3001\u6b8b\u308a\u304c\u6b63\u306e\u30d8\u30ea\u30b7\u30c6\u30a3\u3092\u6301\u3064\u69cb\u6210\uff09\u304c\u3001\u7c92\u5b50\u904b\u52d5\u304c<strong>\u534a\u5171\u7dda\u4f53\u5236<\/strong>\u3068\u547c\u3070\u308c\u308b\u7279\u6b8a\u306a\u30a2\u30e9\u30a4\u30e1\u30f3\u30c8\u3092\u6e80\u305f\u3059\u5834\u5408\u306b\u6d88\u6ec5\u3057\u306a\u3044\u3053\u3068\u3092\u767a\u898b\u3057\u307e\u3057\u305f\u3002<\/p>\n<p>\u3053\u306e\u7814\u7a76\u306e\u7279\u7b46\u3059\u3079\u304d\u70b9\u306f\u3001<strong>GPT-5.2 Pro<\/strong>\u304c\u3053\u306e\u6570\u5b66\u7684\u7d50\u679c\u306e\u767a\u898b\u306b\u8ca2\u732e\u3057\u305f\u3053\u3068\u3067\u3059\u3002\u4ee5\u524d\u306e\u30b0\u30eb\u30fc\u30aa\u30f3\u306b\u95a2\u3059\u308b\u8ad6\u6587\u3092\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u3068\u3057\u3066\u4e0e\u3048\u3089\u308c\u305f<strong>GPT-5.2 Pro<\/strong>\u306f\u3001\u4eba\u9593\u304c\u5c0e\u304d\u51fa\u3059\u306e\u306b\u304b\u306a\u308a\u306e\u6642\u9593\u3092\u8981\u3059\u308b\u3067\u3042\u308d\u3046\u91cf\u5b50\u91cd\u529b\u306b\u304a\u3051\u308b\u5bfe\u5fdc\u3059\u308b\u632f\u5e45\u3092\u69cb\u7bc9\u3059\u308b\u3088\u3046\u6c42\u3081\u3089\u308c\u307e\u3057\u305f\u3002\u30e2\u30c7\u30eb\u306f\u3001<strong>\u6307\u5411\u6027\u884c\u5217\u6728\u5b9a\u7406<\/strong>\u3068\u3044\u3046\u9a5a\u304f\u3079\u304d\u624b\u6cd5\u3092\u7528\u3044\u3066\u554f\u984c\u3092\u89e3\u6c7a\u3057\u3001\u8ad6\u6587\u306e\u512a\u308c\u305f\u4e88\u5099\u8349\u7a3f\u3082\u4f5c\u6210\u3057\u307e\u3057\u305f\u3002<\/p>\n<p>\u3053\u306e\u6210\u679c\u306f\u3001\u91cf\u5b50\u529b\u5b66\u3068<strong>\u30a2\u30a4\u30f3\u30b7\u30e5\u30bf\u30a4\u30f3<\/strong>\u306e\u4e00\u822c\u76f8\u5bfe\u6027\u7406\u8ad6\u3092\u8abf\u548c\u3055\u305b\u308b\u3068\u3044\u3046\u4e2d\u5fc3\u7684\u306a\u554f\u984c\u3078\u306e\u5c0f\u3055\u306a\u4e00\u6b69\u3067\u3042\u308a\u3001<strong>\u30ed\u30b8\u30e3\u30fc\u30fb\u30da\u30f3\u30ed\u30fc\u30ba<\/strong>\u306b\u3088\u3063\u3066\u534a\u4e16\u7d00\u524d\u306b\u767a\u898b\u3055\u308c\u305f\u7121\u9650\u6b21\u5143\u306e\u300cw-(1+\u221e)\u300d\u5bfe\u79f0\u6027\u3092\u91cd\u529b\u5b50\u306b\u9069\u7528\u3059\u308b\u65b9\u6cd5\u3092\u793a\u3057\u3066\u3044\u307e\u3059\u3002\u3053\u306e\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u306f\u3001AI\u652f\u63f4\u63a8\u8ad6\u304c\u5f93\u6765\u306e\u6570\u5b66\u7684\u691c\u8a3c\u3068\u79d1\u5b66\u7684\u53b3\u5bc6\u3055\u3092\u7dad\u6301\u3057\u3064\u3064\u3001\u7406\u8ad6\u7814\u7a76\u306b\u3069\u306e\u3088\u3046\u306b\u8ca2\u732e\u3067\u304d\u308b\u304b\u3092\u793a\u3059\u3082\u306e\u3067\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/openai.com\/index\/extending-single-minus-amplitudes-to-gravitons\/\" target=\"_blank\" rel=\"noopener\">OpenAI \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> GPT-5.2 Pro significantly accelerated the discovery of a new mathematical result in quantum gravity, demonstrating AI's potential to assist in complex theoretical physics research by extending scattering amplitude theories to gravitons and revealing previously unknown interactions.<\/li>\n<li><strong>\u8457\u8005:<\/strong> Alex Lupsasca<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p><strong>OpenAI<\/strong> researchers have published new findings on scattering amplitudes in quantum gravity, extending previous results for gluons to gravitons. This work demonstrates that a class of graviton interactions, long assumed to vanish, can in fact arise under specific kinematic conditions.<\/p>\n<p>The paper, titled \u201cSingle-minus graviton tree amplitudes are nonzero,\u201d authored by <strong>Alfredo Guevara<\/strong>, <strong>Alexandru Lupsasca<\/strong>, <strong>David Skinner<\/strong>, <strong>Andrew Strominger<\/strong>, and <strong>Kevin Weil<\/strong>, reveals that single-minus amplitudes (configurations where one particle has negative helicity and the rest have positive helicity) do not vanish when particle momenta satisfy a special alignment known as the <strong>half-collinear regime<\/strong>.<\/p>\n<p>A significant aspect of this research is the contribution of <strong>GPT-5.2 Pro<\/strong> in discovering this new mathematical result. Provided with the earlier gluon paper as context, <strong>GPT-5.2 Pro<\/strong> was tasked with constructing the corresponding amplitudes in quantum gravity, a derivation that would have taken human authors considerable time. The model solved the problem using a surprising technique, the <strong>directed matrix-tree theorem<\/strong>, and produced an excellent preliminary draft of the paper.<\/p>\n<p>This achievement represents a small step towards reconciling quantum mechanics with <strong>Einstein's<\/strong> theory of general relativity, illustrating how the infinite-dimensional \u201cw-(1+\u221e)\u201d symmetry, discovered by <strong>Roger Penrose<\/strong>, acts on gravitons. The project highlights how AI-assisted reasoning can contribute to theoretical research while upholding conventional standards of mathematical verification and scientific rigor.<\/p>\n<\/blockquote>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">Luma\u304c\u65b0\u3057\u3044\u300cUnified Intelligence\u300d\u30e2\u30c7\u30eb\u3092\u642d\u8f09\u3057\u305f\u30af\u30ea\u30a8\u30a4\u30c6\u30a3\u30d6AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3092\u767a\u8868<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> EXCLUSIVE: Luma launches creative AI agents powered by its new \u2018Unified Intelligence\u2019 models | TechCrunch<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p>AI\u30d3\u30c7\u30aa\u751f\u6210\u30b9\u30bf\u30fc\u30c8\u30a2\u30c3\u30d7\u306e<strong>Luma<\/strong>\u306f\u3001\u65b0\u3057\u3044\u300c<strong>Unified Intelligence<\/strong>\u300d\u30e2\u30c7\u30eb\u3092\u642d\u8f09\u3057\u305f\u30af\u30ea\u30a8\u30a4\u30c6\u30a3\u30d6<strong>AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8<\/strong>\u300c<strong>Luma Agents<\/strong>\u300d\u3092\u767a\u8868\u3057\u307e\u3057\u305f\u3002\u3053\u308c\u3089\u306e\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306f\u3001\u30c6\u30ad\u30b9\u30c8\u3001\u753b\u50cf\u3001\u30d3\u30c7\u30aa\u3001\u30aa\u30fc\u30c7\u30a3\u30aa\u306b\u308f\u305f\u308b\u30a8\u30f3\u30c9\u30c4\u30fc\u30a8\u30f3\u30c9\u306e\u30af\u30ea\u30a8\u30a4\u30c6\u30a3\u30d6\u4f5c\u696d\u3092\u51e6\u7406\u3059\u308b\u3088\u3046\u306b\u8a2d\u8a08\u3055\u308c\u3066\u304a\u308a\u3001\u5e83\u544a\u4ee3\u7406\u5e97\u3001\u30de\u30fc\u30b1\u30c6\u30a3\u30f3\u30b0\u30c1\u30fc\u30e0\u3001\u30c7\u30b6\u30a4\u30f3\u30b9\u30bf\u30b8\u30aa\u3001\u4f01\u696d\u5411\u3051\u306b\u65b0\u305f\u306a\u50cd\u304d\u65b9\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p><strong>Luma Agents<\/strong>\u306f\u3001\u540c\u793e\u306e<strong>Uni-1\u30e2\u30c7\u30eb<\/strong>\u3092\u57fa\u76e4\u3068\u3057\u3066\u304a\u308a\u3001\u3053\u308c\u306f<strong>Unified Intelligence<\/strong>\u30d5\u30a1\u30df\u30ea\u30fc\u306e\u6700\u521d\u306e\u30e2\u30c7\u30eb\u3067\u3059\u3002<strong>Uni-1<\/strong>\u306f\u3001\u30aa\u30fc\u30c7\u30a3\u30aa\u3001\u30d3\u30c7\u30aa\u3001\u753b\u50cf\u3001\u8a00\u8a9e\u3001\u7a7a\u9593\u63a8\u8ad6\u3067\u8a13\u7df4\u3055\u308c\u3066\u304a\u308a\u3001\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u300c\u8a00\u8a9e\u3067\u8003\u3048\u3001\u30d4\u30af\u30bb\u30eb\u3084\u753b\u50cf\u3067\u60f3\u50cf\u3057\u30ec\u30f3\u30c0\u30ea\u30f3\u30b0\u3059\u308b\u300d\u3053\u3068\u3092\u53ef\u80fd\u306b\u3057\u307e\u3059\u3002\u3053\u306e\u30a2\u30d7\u30ed\u30fc\u30c1\u306f\u3001\u8a00\u8a9e\u3001\u30d3\u30b8\u30e7\u30f3\u3001\u751f\u6210\u306e\u305f\u3081\u306e\u500b\u5225\u306e\u30e2\u30c7\u30eb\u3092\u9023\u7d50\u3059\u308b\u306e\u3067\u306f\u306a\u304f\u3001\u5358\u4e00\u306e\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30eb\u63a8\u8ad6\u30b7\u30b9\u30c6\u30e0\u3067\u7406\u89e3\u3068\u751f\u6210\u3092\u884c\u3046\u3068\u3044\u3046\u3001\u696d\u754c\u306e\u4e00\u822c\u7684\u306a\u30a2\u30d7\u30ed\u30fc\u30c1\u304b\u3089\u306e\u8131\u5374\u3092\u610f\u5473\u3057\u307e\u3059\u3002<\/p>\n<p>\u3053\u308c\u3089\u306e\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306f\u3001\u8907\u6570\u306eAI\u30b7\u30b9\u30c6\u30e0\u3092\u8abf\u6574\u3057\u3001<strong>Luma<\/strong>\u306e<strong>Ray 3.14<\/strong>\u3001<strong>Google Veo 3<\/strong>\u3001<strong>ByteDance Seedream<\/strong>\u3001<strong>ElevenLabs<\/strong>\u306e\u97f3\u58f0\u30e2\u30c7\u30eb\u306a\u3069\u3001\u4ed6\u306eAI\u30e2\u30c7\u30eb\u3068\u9023\u643a\u3057\u3066\u52d5\u4f5c\u3057\u307e\u3059\u3002<strong>Luma<\/strong>\u306f\u3001\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u81ea\u8eab\u306e\u51fa\u529b\u3092\u8a55\u4fa1\u3057\u3001\u53cd\u5fa9\u7684\u306a\u81ea\u5df1\u6279\u8a55\u30d7\u30ed\u30bb\u30b9\u3092\u901a\u3058\u3066\u6539\u5584\u3067\u304d\u308b\u3068\u8ff0\u3079\u3066\u3044\u307e\u3059\u3002\u30c7\u30e2\u30f3\u30b9\u30c8\u30ec\u30fc\u30b7\u30e7\u30f3\u3067\u306f\u3001200\u8a9e\u306e\u30d6\u30ea\u30fc\u30d5\u3068\u753b\u50cf\u304b\u3089\u69d8\u3005\u306a\u5e83\u544a\u30ad\u30e3\u30f3\u30da\u30fc\u30f3\u306e\u30a2\u30a4\u30c7\u30a2\u3092\u751f\u6210\u3057\u30011500\u4e07\u30c9\u30eb\u306e\u5e83\u544a\u30ad\u30e3\u30f3\u30da\u30fc\u30f3\u3092\u308f\u305a\u304b40\u6642\u9593\u3067\u7570\u306a\u308b\u56fd\u5411\u3051\u306b\u30ed\u30fc\u30ab\u30e9\u30a4\u30ba\u3067\u304d\u308b\u3053\u3068\u304c\u793a\u3055\u308c\u307e\u3057\u305f\u3002<\/p>\n<p><strong>Luma<\/strong>\u306eCEO\u517c\u5171\u540c\u5275\u8a2d\u8005\u3067\u3042\u308b<strong>Amit Jain<\/strong>\u306f\u3001\u300c\u9867\u5ba2\u306f\u30c4\u30fc\u30eb\u3092\u8cfc\u5165\u3057\u3066\u3044\u308b\u306e\u3067\u306f\u306a\u304f\u3001\u30d3\u30b8\u30cd\u30b9\u306e\u3084\u308a\u65b9\u3092\u5909\u3048\u3066\u3044\u308b\u300d\u3068\u8ff0\u3079\u3066\u304a\u308a\u3001<strong>Luma Agents<\/strong>\u304c\u30af\u30ea\u30a8\u30a4\u30c6\u30a3\u30d6\u30c1\u30fc\u30e0\u304c\u30c4\u30fc\u30eb\u3092\u8abf\u6574\u3059\u308b\u306e\u3067\u306f\u306a\u304f\u3001\u5473\u899a\u3001\u65b9\u5411\u6027\u3001\u6226\u7565\u306b\u96c6\u4e2d\u3067\u304d\u308b\u3088\u3046\u306b\u3059\u308b\u30b3\u30e9\u30dc\u30ec\u30fc\u30bf\u30fc\u3067\u3042\u308b\u3068\u5f37\u8abf\u3057\u3066\u3044\u307e\u3059\u3002\u3053\u306e\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u306f\u3001<strong>Publicis Groupe<\/strong>\u3084<strong>Serviceplan Group<\/strong>\u306a\u3069\u306e\u30b0\u30ed\u30fc\u30d0\u30eb\u30a8\u30f3\u30bf\u30fc\u30d7\u30e9\u30a4\u30ba\u30d1\u30fc\u30c8\u30ca\u30fc\u306b\u3059\u3067\u306b\u5c55\u958b\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/techcrunch.com\/2026\/03\/05\/exclusive-luma-launches-creative-ai-agents-powered-by-its-new-unified-intelligence-models\/\" target=\"_blank\" rel=\"noopener\">TechCrunch \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> Luma's 'Unified Intelligence' models power new AI agents capable of end-to-end creative work across multiple modalities, streamlining complex workflows and enabling autonomous, self-critiquing content generation for enterprises.<\/li>\n<li><strong>\u8457\u8005:<\/strong> Rebecca Bellan<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p><strong>Luma<\/strong>, an AI video generation startup, has launched <strong>Luma Agents<\/strong>, creative <strong>AI agents<\/strong> powered by its new '<strong>Unified Intelligence<\/strong>' models. These agents are designed to handle end-to-end creative work across text, image, video, and audio, proposing a new way of working for advertising agencies, marketing teams, design studios, and enterprises.<\/p>\n<p><strong>Luma Agents<\/strong> are built on the company's <strong>Uni-1 model<\/strong>, the first in its <strong>Unified Intelligence<\/strong> family. <strong>Uni-1<\/strong> has been trained on audio, video, image, language, and spatial reasoning, enabling the agents to 'think in language and imagine and render in pixels or images.' This approach signifies a departure from the industry's prevailing method of chaining together separate models for language, vision, and generation, instead utilizing a single multimodal reasoning system for understanding and generation.<\/p>\n<p>These agents coordinate multiple AI systems and work with other AI models, including <strong>Luma's Ray 3.14<\/strong>, <strong>Google Veo 3<\/strong>, <strong>ByteDance Seedream<\/strong>, and <strong>ElevenLabs<\/strong>' voice models. <strong>Luma<\/strong> states that the agents can evaluate and refine their own outputs through an iterative self-critique process. Demonstrations showed the agents generating various ad campaign ideas from a 200-word brief and an image, and localizing a $15 million ad campaign for different countries in just 40 hours.<\/p>\n<p><strong>Amit Jain<\/strong>, CEO and co-founder of <strong>Luma<\/strong>, emphasized that customers are not just buying a tool but are redoing how business is done, positioning <strong>Luma Agents<\/strong> as collaborators that allow creative teams to focus on taste, direction, and strategy rather than orchestrating tools. The platform has already been rolled out with global enterprise partners such as <strong>Publicis Groupe<\/strong> and <strong>Serviceplan Group<\/strong>.<\/p>\n<\/blockquote>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>OpenAI\u306eGPT-5.4\u306e\u9032\u5316\u3001Luma AI\u306e\u30af\u30ea\u30a8\u30a4\u30c6\u30a3\u30d6AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3001\u305d\u3057\u3066GPT-5.2 Pro\u304c\u8ca2\u732e\u3057\u305f\u91cf\u5b50\u91cd\u529b\u7814\u7a76\u306e\u6700\u65b0\u52d5\u5411\u3092\u6df1\u6398\u308a\u3002AI\u306e\u591a\u89d2\u7684\u306a\u9032\u6b69\u3092\u89e3\u8aac\u3002<\/p>\n","protected":false},"author":1,"featured_media":855,"comment_status":"","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"vkexunit_cta_each_option":"","footnotes":""},"categories":[3],"tags":[8,17,16,230,57,45,26,15,229],"class_list":{"0":"post-949","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","6":"hentry","7":"category-columns","8":"tag-ai","10":"tag-llm","11":"tag-luma-ai","12":"tag-openai","14":"tag-26","16":"tag-229"},"_links":{"self":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts\/949","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/comments?post=949"}],"version-history":[{"count":0,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts\/949\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/media\/855"}],"wp:attachment":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/media?parent=949"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/categories?post=949"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/tags?post=949"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}