{"id":79,"date":"2026-06-11T11:49:01","date_gmt":"2026-06-11T02:49:01","guid":{"rendered":"https:\/\/www.wetch.co.jp\/ai\/?p=79"},"modified":"2026-06-11T11:49:01","modified_gmt":"2026-06-11T02:49:01","slug":"mistral-ai-now-summit-2026%ef%bc%86nvidia-nemotron-3-ultra%ef%bc%9a6%e6%9c%88%e3%81%ae%e3%82%aa%e3%83%bc%e3%83%97%e3%83%b3%e3%82%bd%e3%83%bc%e3%82%b9llm%e6%9c%80%e5%89%8d%e7%b7%9a","status":"publish","type":"post","link":"https:\/\/www.wetch.co.jp\/ai\/2026\/06\/11\/mistral-ai-now-summit-2026%ef%bc%86nvidia-nemotron-3-ultra%ef%bc%9a6%e6%9c%88%e3%81%ae%e3%82%aa%e3%83%bc%e3%83%97%e3%83%b3%e3%82%bd%e3%83%bc%e3%82%b9llm%e6%9c%80%e5%89%8d%e7%b7%9a\/","title":{"rendered":"Mistral AI Now Summit 2026\uff06NVIDIA Nemotron 3 Ultra\uff1a6\u6708\u306e\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9LLM\u6700\u524d\u7dda"},"content":{"rendered":"<p><em>2026\u5e746\u6708\u3001\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9\u30fb\u30ed\u30fc\u30ab\u30ebLLM\u306e\u4e16\u754c\u306f\u5927\u304d\u304f\u52d5\u3044\u305f\u3002\u30d5\u30e9\u30f3\u30b9\u767a\u306e<a href=\"https:\/\/mistral.ai\/\" target=\"_blank\" rel=\"noopener\">Mistral AI<\/a>\u304c\u300cAI Now Summit 2026\u300d\u3067\u8907\u6570\u30e2\u30c7\u30eb\u3092\u4e00\u6589\u516c\u958b\u3057\u3001NVIDIA\u306f550B\u30d1\u30e9\u30e1\u30fc\u30bf\u306e<a href=\"https:\/\/developer.nvidia.com\/blog\/nvidia-nemotron-3-ultra-powers-faster-more-efficient-reasoning-for-long-running-agents\" target=\"_blank\" rel=\"noopener\">Nemotron 3 Ultra<\/a>\u3092\u30aa\u30fc\u30d7\u30f3\u30a6\u30a7\u30a4\u30c8\u3067\u30ea\u30ea\u30fc\u30b9\u3002\u3055\u3089\u306bAMD GPU\u4e0a\u3067\u306e\u30ed\u30fc\u30ab\u30ebLLM\u74b0\u5883\u6574\u5099\u3082\u9032\u307f\u3001\u30a8\u30f3\u30b8\u30cb\u30a2\u306e\u9078\u629e\u80a2\u304c\u304b\u3064\u3066\u306a\u304f\u5e83\u304c\u3063\u3066\u3044\u308b\u3002<\/em><\/p>\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/www.wetch.co.jp\/ai\/wp-content\/uploads\/2026\/06\/ill-oss.png\" alt=\"\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9LLM\u306e\u6700\u65b0\u52d5\u5411\"\/><\/figure>\n<h2>Mistral AI Now Summit 2026\uff1a\u4e00\u6589\u306b5\u3064\u306e\u767a\u8868<\/h2>\n<p>Mistral AI\u306f2026\u5e746\u6708\u958b\u50ac\u306e\u81ea\u793e\u30a4\u30d9\u30f3\u30c8\u300c<a href=\"https:\/\/mistral.ai\/news\/ai-now-summit-2026\" target=\"_blank\" rel=\"noopener\">AI Now Summit 2026<\/a>\u300d\u3067\u3001\u30e2\u30c7\u30eb\u30fb\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u30fb\u7523\u696dAI\u306b\u308f\u305f\u308b\u8907\u6570\u306e\u767a\u8868\u3092\u884c\u3063\u305f\u3002<\/p>\n<h3>Devstral 2 \uff06 Devstral Small 2\uff1a\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u306e\u65b0\u57fa\u6e96<\/h3>\n<p><a href=\"https:\/\/mistral.ai\/news\/devstral-2-vibe-cli\" target=\"_blank\" rel=\"noopener\">Devstral 2<\/a>\u306f\u3001\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u578b\u30bd\u30d5\u30c8\u30a6\u30a7\u30a2\u30a8\u30f3\u30b8\u30cb\u30a2\u30ea\u30f3\u30b0\u5411\u3051\u306b\u8a2d\u8a08\u3055\u308c\u305f123B\u30d1\u30e9\u30e1\u30fc\u30bf\u306e\u30e2\u30c7\u30eb\u3067\u3001\u30e9\u30a4\u30bb\u30f3\u30b9\u306f<strong>Apache 2.0<\/strong>\u3002<a href=\"https:\/\/swebench.com\/\" target=\"_blank\" rel=\"noopener\">SWE-bench Verified<\/a>\u3067<strong>72.2%<\/strong>\u3092\u8a18\u9332\u3057\u3001256K\u30c8\u30fc\u30af\u30f3\u306e\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u30a6\u30a3\u30f3\u30c9\u30a6\u3092\u6301\u3064\u3002Mistral\u306f\u300cGPT-5.2 Codex \u3068\u6bd4\u3079\u30667\u500d\u30b3\u30b9\u30c8\u52b9\u7387\u304c\u9ad8\u304f\u3001DeepSeek V3.2 \u3088\u308a5\u500d\u5c0f\u3055\u3044\u300d\u3068\u8aac\u660e\u3057\u3066\u3044\u308b\u3002<\/p>\n<p>\u5c0f\u578b\u7248\u306e<strong>Devstral Small 2<\/strong>\uff0824B\u30d1\u30e9\u30e1\u30fc\u30bf\uff09\u306fSWE-bench Verified\u3067<strong>68%<\/strong>\u3092\u9054\u6210\u3057\u3066\u304a\u308a\u3001RTX 4090 1\u679a\u307e\u305f\u306fRAM 32GB\u306eMac\u3067\u52d5\u4f5c\u3059\u308b\u3002\u753b\u50cf\u5165\u529b\u306b\u3082\u5bfe\u5fdc\u3057\u3001\u81ea\u5df1\u30db\u30b9\u30c8\u74b0\u5883\u306b\u9069\u3057\u305f\u9078\u629e\u80a2\u3068\u306a\u3063\u3066\u3044\u308b\u3002<\/p>\n<p>\u3042\u308f\u305b\u3066\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9\u306e\u30bf\u30fc\u30df\u30ca\u30eb\u5411\u3051\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u30a2\u30b7\u30b9\u30bf\u30f3\u30c8 <a href=\"https:\/\/mistral.ai\/products\/vibe\/code\/\" target=\"_blank\" rel=\"noopener\">Vibe CLI<\/a>\uff08Apache 2.0\uff09\u3082\u516c\u958b\u3055\u308c\u305f\u3002VS Code\u5411\u3051\u62e1\u5f35\u6a5f\u80fd\u3082\u63d0\u4f9b\u3055\u308c\u3066\u304a\u308a\u3001Devstral 2\u3092\u30d0\u30c3\u30af\u30a8\u30f3\u30c9\u3068\u3057\u3066\u5229\u7528\u3067\u304d\u308b\u3002<\/p>\n<p>\u51fa\u5178: <a href=\"https:\/\/pinggy.io\/blog\/best_open_source_self_hosted_llms_for_coding\" target=\"_blank\" rel=\"noopener\">Pinggy &#8211; Best Open Source Self-Hosted LLMs for Coding in 2026<\/a><\/p>\n<h3>Vibe Agent\uff1aWork \u3068 Code \u306e\u7d71\u5408\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8<\/h3>\n<p><a href=\"https:\/\/mistral.ai\/news\/vibe-agent\" target=\"_blank\" rel=\"noopener\">Vibe<\/a>\u306f\u3001\u9577\u6642\u9593\u306b\u308f\u305f\u308b\u30bf\u30b9\u30af\u51e6\u7406\u3092\u60f3\u5b9a\u3057\u305f\u7d71\u5408\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3067\u3001\u300cWork \u30e2\u30fc\u30c9\u300d\uff08\u8907\u96d1\u306a\u696d\u52d9\u30bf\u30b9\u30af\uff09\u3068\u300cCode \u30e2\u30fc\u30c9\u300d\uff08\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\uff09\u306e2\u3064\u3092\u63d0\u4f9b\u3059\u308b\u3002<a href=\"https:\/\/chat.mistral.ai\/\" target=\"_blank\" rel=\"noopener\">Le Chat<\/a>\u4e0a\u3067\u52d5\u4f5c\u3057\u3001\u6709\u6599\u30d7\u30e9\u30f3\u306fLe Chat Pro\uff08\u6708\u984d$15\u301c\uff09\u304b\u3089\u5229\u7528\u53ef\u80fd\u3002\u30ea\u30e2\u30fc\u30c8\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u6a5f\u80fd\u3082\u8ffd\u52a0\u3055\u308c\u305f\u3002<\/p>\n<p>\u51fa\u5178: <a href=\"https:\/\/mistral.ai\/news\/vibe-agent\" target=\"_blank\" rel=\"noopener\">Mistral AI &#8211; Vibe gets to work.<\/a><\/p>\n<h3>Mistral Medium 3.5 \u3068 Mistral Small 4<\/h3>\n<p><a href=\"https:\/\/mistral.ai\/news\/vibe-remote-agents-mistral-medium-3-5\" target=\"_blank\" rel=\"noopener\">Mistral Medium 3.5<\/a>\u306fVibe \u30ea\u30e2\u30fc\u30c8\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3068\u3068\u3082\u306b\u767a\u8868\u3055\u308c\u305f\u30df\u30c3\u30c9\u30c6\u30a3\u30a2\u30e2\u30c7\u30eb\u3002<strong>Mistral Small 4<\/strong>\uff08119B-A6B \u306e MoE\uff09\u3082\u30ea\u30ea\u30fc\u30b9\u3055\u308c\u3001Artificial Analysis\u306eCoding Index\u3067\u4e3b\u8981\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9\u30e2\u30c7\u30eb\u3068\u7af6\u5408\u3057\u3066\u3044\u308b\u3002<\/p>\n<p>\u51fa\u5178: <a href=\"https:\/\/mistral.ai\/news\" target=\"_blank\" rel=\"noopener\">Mistral AI News<\/a><\/p>\n<h3>Physics AI\uff1a\u7523\u696d\u30a8\u30f3\u30b8\u30cb\u30a2\u30ea\u30f3\u30b0\u3078\u306e\u672c\u683c\u53c2\u5165<\/h3>\n<p>Mistral\u306f2026\u5e745\u6708\u306b\u30a8\u30f3\u30b8\u30cb\u30a2\u30ea\u30f3\u30b0AI\u4f01\u696d<a href=\"https:\/\/www.emmi.ai\/news\/mistral-ai-acquires-emmi-ai\" target=\"_blank\" rel=\"noopener\">Emmi AI<\/a>\uff08\u30aa\u30fc\u30b9\u30c8\u30ea\u30a2\u30fb\u30ea\u30f3\u30c4\uff09\u3092\u8cb7\u53ce\u3057\u3001\u7269\u7406\u30b7\u30b9\u30c6\u30e0\u306e\u6319\u52d5\u3092\u4e88\u6e2c\u3059\u308b\u300cPhysics AI\u300d\u30e2\u30c7\u30eb\u306e\u958b\u767a\u3092\u672c\u683c\u5316\u3002\u822a\u7a7a\u5b87\u5b99\u30fb\u81ea\u52d5\u8eca\u30fb\u534a\u5c0e\u4f53\u5206\u91ce\u3067\u306e\u7523\u696d\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\u52a0\u901f\u3092\u76ee\u6307\u3059\u3002BMW\u30b0\u30eb\u30fc\u30d7\u3068\u3082\u300cLarge Industry Model\u300d\u69cb\u60f3\u3067\u9023\u643a\u3057\u3001\u885d\u7a81\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\u7b49\u306e\u8907\u96d1\u306a\u30a8\u30f3\u30b8\u30cb\u30a2\u30ea\u30f3\u30b0\u30bf\u30b9\u30af\u3078\u306e\u9069\u7528\u3092\u9032\u3081\u3066\u3044\u308b\u3002<\/p>\n<p>\u51fa\u5178: <a href=\"https:\/\/venturebeat.com\/technology\/mistral-ai-launches-vibe-expands-into-industrial-ai-and-announces-data-center-push-to-challenge-openai\" target=\"_blank\" rel=\"noopener\">VentureBeat &#8211; Mistral AI launches Vibe, expands into industrial AI<\/a><\/p>\n<h3>NVIDIA \u3068\u306e\u63d0\u643a\u3067\u30aa\u30fc\u30d7\u30f3\u30d5\u30ed\u30f3\u30c6\u30a3\u30a2\u30e2\u30c7\u30eb\u3092\u52a0\u901f<\/h3>\n<p>Mistral\u306f<a href=\"https:\/\/mistral.ai\/news\/mistral-ai-and-nvidia-partner-to-accelerate-open-frontier-models\" target=\"_blank\" rel=\"noopener\">NVIDIA \u3068\u306e\u63d0\u643a<\/a>\u3082\u767a\u8868\u3002NVIDIA\u306e\u30a4\u30f3\u30d5\u30e9\u4e0a\u3067\u30aa\u30fc\u30d7\u30f3\u30a6\u30a7\u30a4\u30c8\u30e2\u30c7\u30eb\u306e\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u30fb\u63a8\u8ad6\u3092\u6700\u9069\u5316\u3057\u3001\u30a8\u30f3\u30bf\u30fc\u30d7\u30e9\u30a4\u30ba\u5411\u3051\u5c55\u958b\u3092\u5f37\u5316\u3059\u308b\u65b9\u91dd\u3060\u3002<\/p>\n<hr \/>\n<h2>NVIDIA Nemotron 3 \u30d5\u30a1\u30df\u30ea\u30fc\uff1aUltra\u30fbSuper\u30fbNano<\/h2>\n<p>NVIDIA\u304c2026\u5e746\u6708\u306b\u30ea\u30ea\u30fc\u30b9\u3057\u305f<a href=\"https:\/\/developer.nvidia.com\/blog\/nvidia-nemotron-3-ultra-powers-faster-more-efficient-reasoning-for-long-running-agents\" target=\"_blank\" rel=\"noopener\">Nemotron 3 Ultra<\/a>\u306f\u3001OSS LLM\u754c\u9688\u3067\u300c\u4eca\u9031\u6700\u91cd\u8981\u30ea\u30ea\u30fc\u30b9\u300d\u3068\u8a55\u3055\u308c\u308b\u6ce8\u76ee\u4f5c\u3060\u3002<\/p>\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/www.wetch.co.jp\/ai\/wp-content\/uploads\/2026\/06\/photo.webp\" alt=\"AI\u3068\u8a08\u7b97\u306e\u672a\u6765\u3092\u793a\u3059\u30b3\u30f3\u30bb\u30d7\u30c8\u753b\u50cf\"\/><figcaption><small>Photo: <a href=\"https:\/\/www.flickr.com\/photos\/44124348109@N01\/28799579322\" target=\"_blank\" rel=\"noopener\">jurvetson<\/a> \/ CC BY\uff08Openverse\u7d4c\u7531\uff09<\/small><\/figcaption><\/figure>\n<h3>Nemotron 3 Ultra\uff1a550B MoE\u30011M\u30c8\u30fc\u30af\u30f3\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8<\/h3>\n<ul>\n<li><strong>\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\uff1a<\/strong>550B\u30d1\u30e9\u30e1\u30fc\u30bf\u3001\u30a2\u30af\u30c6\u30a3\u30d655B\uff08MoE\uff09\u306e\u30cf\u30a4\u30d6\u30ea\u30c3\u30c9 Mamba-Transformer<\/li>\n<li><strong>\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\uff1a<\/strong>1M\u30c8\u30fc\u30af\u30f3\uff08RULER@1M\u3067\u9577\u6587\u66f8\u53d6\u5f97\u7cbe\u5ea6 91.8 \u3092\u9054\u6210\uff09<\/li>\n<li><strong>\u7cbe\u5ea6\uff1a<\/strong>MMLU 89.1\u3001SWE-Bench Verified \u306730%\u30b3\u30b9\u30c8\u524a\u6e1b\u3092\u5b9f\u73fe<\/li>\n<li><strong>\u30b9\u30eb\u30fc\u30d7\u30c3\u30c8\uff1a<\/strong>NVFP4\u30d0\u30ea\u30a2\u30f3\u30c8\u3067\u7d045\u500d\u306e\u63a8\u8ad6\u901f\u5ea6\u5411\u4e0a<\/li>\n<li><strong>\u30e9\u30a4\u30bb\u30f3\u30b9\uff1a<\/strong>NVIDIA Nemotron Open Model License\uff08\u30aa\u30fc\u30d7\u30f3\u30a6\u30a7\u30a4\u30c8\uff09<\/li>\n<li><strong>\u8a13\u7df4\uff1a<\/strong>Multi-Teacher On-Policy Distillation\uff08MOPD\uff09\u3092\u63a1\u7528<\/li>\n<\/ul>\n<p>\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u5411\u3051\u5b89\u5168\u6a5f\u80fd\u3068\u3057\u3066 <strong>NemoClaw<\/strong>\uff08\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u884c\u52d5\u306e\u5236\u9650\u30fb\u76e3\u8996\uff09\u3068 <strong>OpenShell<\/strong>\uff08\u97f3\u58f0\u5bfe\u5fdc\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\uff09\u3082\u540c\u6642\u516c\u958b\u3055\u308c\u305f\u3002<\/p>\n<h3>Nemotron 3 Super\uff08120B-A12B\uff09\u3068 Nano\uff084B \/ 30B-A3B\uff09<\/h3>\n<p>Ultra\u306b\u5148\u884c\u3057\u3066\u516c\u958b\u3055\u308c\u305f<strong>Nemotron 3 Super<\/strong>\uff08120B-A12B\uff09\u306f1M\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u306b\u5bfe\u5fdc\u3057\u3001OpenHands\u3067\u306eSWE-bench\u8a55\u4fa1\u306760.5%\u3092\u9054\u6210\u3002<strong>Nemotron 3 Nano-4B<\/strong>\u306f\u308f\u305a\u304b5GB\u306eVRAM\/RAM\u3067\u52d5\u4f5c\u3057\u3001<a href=\"https:\/\/unsloth.ai\/docs\/models\/nemotron-3\" target=\"_blank\" rel=\"noopener\">Unsloth<\/a>\u7d4c\u7531\u3067\u30ed\u30fc\u30ab\u30eb\u3067\u306e\u30d5\u30a1\u30a4\u30f3\u30c1\u30e5\u30fc\u30cb\u30f3\u30b0\u3082\u53ef\u80fd\u3060\u3002<\/p>\n<p>\u51fa\u5178: <a href=\"https:\/\/developer.nvidia.com\/blog\/nvidia-nemotron-3-ultra-powers-faster-more-efficient-reasoning-for-long-running-agents\" target=\"_blank\" rel=\"noopener\">NVIDIA Developer Blog &#8211; Nemotron 3 Ultra<\/a><\/p>\n<hr \/>\n<h2>\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u7279\u5316\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9\u30e2\u30c7\u30eb \u6027\u80fd\u6bd4\u8f03\uff082026\u5e745\u6708\u6642\u70b9\uff09<\/h2>\n<p>\u4ee5\u4e0b\u306f LiveBench Coding Average\uff082026\u5e745\u670812\u65e5\u30b9\u30ca\u30c3\u30d7\u30b7\u30e7\u30c3\u30c8\uff09\u306b\u57fa\u3065\u304f\u4e3b\u8981\u30e2\u30c7\u30eb\u306e\u6bd4\u8f03\u3002OSS\u9663\u55b6\u304c\u6025\u901f\u306b\u30af\u30ed\u30fc\u30ba\u30c9\u30e2\u30c7\u30eb\u306b\u8feb\u3063\u3066\u3044\u308b\u3002<\/p>\n<table>\n<thead>\n<tr>\n<th>\u30e2\u30c7\u30eb<\/th>\n<th>\u63d0\u4f9b\u5143<\/th>\n<th>\u7a2e\u5225<\/th>\n<th>LiveBench Coding Avg<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>GPT-5.2 Codex<\/td>\n<td>OpenAI<\/td>\n<td>\u30d7\u30ed\u30d7\u30e9\u30a4\u30a8\u30bf\u30ea<\/td>\n<td>83.62<\/td>\n<\/tr>\n<tr>\n<td>Kimi K2.6 Thinking<\/td>\n<td>Moonshot AI<\/td>\n<td>\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9<\/td>\n<td>78.57<\/td>\n<\/tr>\n<tr>\n<td>DeepSeek V3.2<\/td>\n<td>DeepSeek<\/td>\n<td>\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9<\/td>\n<td>75.69<\/td>\n<\/tr>\n<tr>\n<td>GLM 5.1<\/td>\n<td>Z.AI<\/td>\n<td>\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9<\/td>\n<td>75.37<\/td>\n<\/tr>\n<tr>\n<td>Qwen 3.6 27B<\/td>\n<td>Alibaba<\/td>\n<td>\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9<\/td>\n<td>71.78<\/td>\n<\/tr>\n<tr>\n<td>MiniMax M2.5<\/td>\n<td>MiniMax<\/td>\n<td>\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9<\/td>\n<td>70.70<\/td>\n<\/tr>\n<tr>\n<td><strong>Devstral 2<\/strong><\/td>\n<td><strong>Mistral<\/strong><\/td>\n<td><strong>\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9<\/strong><\/td>\n<td><strong>66.79<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>Devstral 2\u306f\u5358\u4f53\u6027\u80fd\u3067\u4e0a\u4f4d\u30e2\u30c7\u30eb\u306b\u5dee\u304c\u3042\u308b\u3082\u306e\u306e\u3001<strong>Apache 2.0\u30e9\u30a4\u30bb\u30f3\u30b9\u30fb\u81ea\u5df1\u30db\u30b9\u30c8\u53ef\u80fd\u30fb\u30b3\u30b9\u30c8\u52b9\u7387<\/strong>\u306e\u70b9\u3067\u4f01\u696d\u5c0e\u5165\u306b\u5b9f\u7528\u7684\u306a\u9078\u629e\u80a2\u3068\u306a\u3063\u3066\u3044\u308b\u3002<\/p>\n<p>\u51fa\u5178: <a href=\"https:\/\/pinggy.io\/blog\/best_open_source_self_hosted_llms_for_coding\" target=\"_blank\" rel=\"noopener\">Pinggy &#8211; Best Open Source Self-Hosted LLMs for Coding in 2026<\/a>\u3001<a href=\"https:\/\/kilo.ai\/open-source-models\" target=\"_blank\" rel=\"noopener\">Kilo Code &#8211; Best Open-Source Coding Models<\/a><\/p>\n<hr \/>\n<h2>AMD ROCm \u5bfe\u5fdc\u5f37\u5316\uff1aGeForce\u4ee5\u5916\u3067\u3082\u30ed\u30fc\u30ab\u30ebLLM\u304c\u73fe\u5b9f\u7684\u306b<\/h2>\n<p>2026\u5e74\u306e\u30ed\u30fc\u30ab\u30ebLLM\u74b0\u5883\u3067\u306f\u3001AMD GPU\u306e\u9078\u629e\u80a2\u304c\u73fe\u5b9f\u7684\u306b\u306a\u3063\u3066\u304d\u305f\u3002<a href=\"https:\/\/www.mindstudio.ai\/blog\/running-local-ai-amd-rocm-ollama-lm-studio\" target=\"_blank\" rel=\"noopener\">MindStudio \u306e\u30ac\u30a4\u30c9<\/a>\u306b\u3088\u308b\u3068\u3001AMD \u306e ROCm \u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u304c PyTorch\u30fb<a href=\"https:\/\/ollama.com\/\" target=\"_blank\" rel=\"noopener\">Ollama<\/a>\u30fb<a href=\"https:\/\/lmstudio.ai\/\" target=\"_blank\" rel=\"noopener\">LM Studio<\/a>\u30fbComfyUI \u3092\u5b89\u5b9a\u30b5\u30dd\u30fc\u30c8\u3059\u308b\u3088\u3046\u306b\u306a\u308a\u300132GB Radeon GPU \u642d\u8f09\u30de\u30b7\u30f3\u3067\u306e\u5b9f\u7528\u7684\u306a\u30ed\u30fc\u30ab\u30eb\u63a8\u8ad6\u304c\u53ef\u80fd\u306b\u306a\u3063\u3066\u3044\u308b\u3002<\/p>\n<ul>\n<li>ROCm \u304c PyTorch\u30fbOllama\u30fbLM Studio\u30fbComfyUI \u3092\u6b63\u5f0f\u30b5\u30dd\u30fc\u30c8<\/li>\n<li>32GB Radeon GPU\uff08\u4f8b\uff1aRX 7900 XTX\uff09\u3067 30B \u898f\u6a21\u306e\u30e2\u30c7\u30eb\u3092\u5b9f\u884c\u53ef\u80fd<\/li>\n<li>NVIDIA CUDA \u3068\u306e\u4e92\u63db\u6027\u30ec\u30a4\u30e4\u30fc\u3067\u65e2\u5b58\u30b9\u30af\u30ea\u30d7\u30c8\u3092\u307b\u307c\u305d\u306e\u307e\u307e\u5229\u7528\u3067\u304d\u308b\u30b1\u30fc\u30b9\u3082<\/li>\n<li>Nemotron 3 Nano-4B\uff085GB VRAM\uff09\u306e\u3088\u3046\u306a\u8efd\u91cf\u30e2\u30c7\u30eb\u306a\u3089\u5ec9\u4fa1\u306aAMD GPU\u3067\u3082\u52d5\u4f5c<\/li>\n<\/ul>\n<p>\u51fa\u5178: <a href=\"https:\/\/www.mindstudio.ai\/blog\/running-local-ai-amd-rocm-ollama-lm-studio\" target=\"_blank\" rel=\"noopener\">MindStudio &#8211; Running Local AI on AMD: ROCm, Ollama, and LM Studio<\/a><\/p>\n<hr \/>\n<h2>\u4e3b\u306a\u60c5\u5831\u6e90<\/h2>\n<ul>\n<li><a href=\"https:\/\/mistral.ai\/news\/ai-now-summit-2026\" target=\"_blank\" rel=\"noopener\">Mistral AI &#8211; AI Now Summit 2026<\/a><\/li>\n<li><a href=\"https:\/\/mistral.ai\/news\/vibe-agent\" target=\"_blank\" rel=\"noopener\">Mistral AI &#8211; Vibe gets to work.<\/a><\/li>\n<li><a href=\"https:\/\/developer.nvidia.com\/blog\/nvidia-nemotron-3-ultra-powers-faster-more-efficient-reasoning-for-long-running-agents\" target=\"_blank\" rel=\"noopener\">NVIDIA Developer Blog &#8211; Nemotron 3 Ultra<\/a><\/li>\n<li><a href=\"https:\/\/www.emmi.ai\/news\/mistral-ai-acquires-emmi-ai\" target=\"_blank\" rel=\"noopener\">Emmi AI &#8211; Mistral AI Acquires Emmi AI<\/a><\/li>\n<li><a href=\"https:\/\/venturebeat.com\/technology\/mistral-ai-launches-vibe-expands-into-industrial-ai-and-announces-data-center-push-to-challenge-openai\" target=\"_blank\" rel=\"noopener\">VentureBeat &#8211; Mistral AI launches Vibe, expands into industrial AI<\/a><\/li>\n<li><a href=\"https:\/\/pinggy.io\/blog\/best_open_source_self_hosted_llms_for_coding\" target=\"_blank\" rel=\"noopener\">Pinggy &#8211; Best Open Source Self-Hosted LLMs for Coding in 2026<\/a><\/li>\n<li><a href=\"https:\/\/kilo.ai\/open-source-models\" target=\"_blank\" rel=\"noopener\">Kilo Code &#8211; Best Open-Source &amp; Open-Weight Coding Models (2026)<\/a><\/li>\n<li><a href=\"https:\/\/www.mindstudio.ai\/blog\/running-local-ai-amd-rocm-ollama-lm-studio\" target=\"_blank\" rel=\"noopener\">MindStudio &#8211; Running Local AI on AMD: ROCm, Ollama, and LM Studio<\/a><\/li>\n<li><a href=\"https:\/\/unsloth.ai\/docs\/models\/nemotron-3\" target=\"_blank\" rel=\"noopener\">Unsloth &#8211; NVIDIA Nemotron 3 Nano Run Guide<\/a><\/li>\n<\/ul>\n<p><small>\u5404\u30b5\u30fc\u30d3\u30b9\u306f\u63d0\u4f9b\u5143\u306e\u898f\u7d04\u306b\u5f93\u3063\u3066\u3054\u5229\u7528\u304f\u3060\u3055\u3044\u3002\u60c5\u5831\u306f\u5404\u60c5\u5831\u6e90\u306e\u516c\u958b\u6642\u70b9\u306e\u3082\u306e\u3067\u3059\u3002\u53ce\u96c6\u30fb\u8981\u7d04\u306fAI\u3001\u516c\u958b\u5224\u65ad\u306f\u4eba\u3002<\/small><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Mistral AI\u304cAI Now Summit 2026\u3067Vibe Agent\u30fbDevstral 2\u30fbPhysics AI\u306a\u3069\u8907\u6570\u30e2\u30c7\u30eb\u3092\u4e00\u6589\u767a\u8868\u3002NVIDIA\u3082Nemotron 3 Ultra\u3067550B MoE\u30e2\u30c7\u30eb\u3092\u30aa\u30fc\u30d7\u30f3\u30ea\u30ea\u30fc\u30b9\u3002\u30ed\u30fc\u30ab\u30eb\u5b9f\u884c\u74b0\u5883\u3067\u306fAMD ROCm\u5bfe\u5fdc\u3082\u9032\u307f\u3001OSS LLM\u306e\u9078\u629e\u80a2\u304c\u6025\u62e1\u5927\u3057\u305f6\u6708\u306e\u52d5\u5411\u3092\u307e\u3068\u3081\u307e\u3059\u3002<\/p>\n","protected":false},"author":1,"featured_media":58,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"page_builder":"","footnotes":""},"categories":[46],"tags":[],"class_list":["post-79","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-aillm"],"_links":{"self":[{"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/posts\/79","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/comments?post=79"}],"version-history":[{"count":1,"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/posts\/79\/revisions"}],"predecessor-version":[{"id":81,"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/posts\/79\/revisions\/81"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/media\/58"}],"wp:attachment":[{"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/media?parent=79"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/categories?post=79"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wetch.co.jp\/ai\/wp-json\/wp\/v2\/tags?post=79"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}