{"id":9700,"date":"2026-11-20T09:43:59","date_gmt":"2026-11-20T09:43:59","guid":{"rendered":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model"},"modified":"2026-11-20T09:43:59","modified_gmt":"2026-11-20T09:43:59","slug":"tencent-hy4-open-weights-770b-multimodal-model","status":"publish","type":"post","link":"https:\/\/i10x.ai\/news\/tencent-hy4-open-weights-770b-multimodal-model","title":{"rendered":"Tencent Hy4 Open Weights: The 770B Multimodal Flagship Is Now Free to Download, With a 1M Token Context"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">Tencent has released the open weights of Hy4, its new flagship AI model. Hy4 is a mixture-of-experts model with 770 billion total parameters, of which 49 billion are active for each token, and it reads text, images and video within a context window of 1 million tokens. The instruct model, an FP8 version and the pre-trained base model are now downloadable, two days after Tencent opened up the smaller Hy4 Flash.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\u26a1 Quick Take<\/h2>\n\n\n\n<ul class=\"wp-block-list\"><li><strong>What&#8217;s new:<\/strong> Hy4, Hy4-FP8 and Hy4 base are open-weight on Hugging Face, ModelScope and GitCode, announced by Tencent Hy on 20 November 2026.<\/li><li><strong>Size:<\/strong> 770B total parameters, 49B active for text plus 1.23B for vision, 256 routed experts per layer and a 1,048,576 token context.<\/li><li><strong>New skills:<\/strong> Tencent says Hy4 adds vision and computer use across command-line and graphical interfaces for the first time.<\/li><li><strong>Hardware:<\/strong> the FP8 checkpoint is about 786 GB and runs on eight H200 or B200 GPUs, or four B300 or GB300 GPUs.<\/li><li><strong>License:<\/strong> the Tencent Hy License, derived from Apache 2.0, requires a separate license for large model-as-a-service providers.<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">In plain English<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Tencent, the Chinese company behind WeChat, has published the full &#8220;brain&#8221; of its best AI model so anyone can download it and run it on their own servers. It can read pictures and video as well as text, and it can operate a computer. It is very large, so it needs a rack of top-end graphics chips, but companies can now use it without sending data to Tencent.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What happened<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The Tencent Hy team posted on X on 20 November: &#8220;Today, we&#8217;re releasing the open weights for Hy4.&#8221; It called the model &#8220;Our natively multimodal flagship MoE (770B total \/ 49B active, 1M context)&#8221; and published a system card with safety, security and alignment evaluations alongside the weights, including a look at cyber and biological risks.<\/p>\n\n\n\n<blockquote class=\"twitter-tweet i10x-x-embed\"><p lang=\"en\" dir=\"ltr\">Today, we&#x27;re releasing the open weights for Hy4.<br><br>Our natively multimodal flagship MoE (770B total \/ 49B active, 1M context).<br><br>Alongside the Hy4 System Card with a comprehensive safety evaluation report.<\/p>Tencent Hy (@TencentHunyuan) \u00b7 <a href=\"https:\/\/x.com\/TencentHunyuan\/status\/2123625549147070538\" target=\"_blank\" rel=\"noreferrer noopener\">November 20, 2026 \u00b7 View on X<\/a><\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">vLLM and SGLang shipped day-one support, and Together AI made Hy4 available on its platform the same day. Tencent&#8217;s Hy4 page says the model is also on Tencent Cloud and OpenRouter and inside its own products CodeBuddy, WorkBuddy, Yuanbao and ima, at an API price of $0.834 per million input tokens, $2.501 per million output tokens and $0.042 per million cached input tokens.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Background<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Hy4 follows a preview version and the lighter Hy4 Flash, which we covered when <a href=\"https:\/\/i10x.ai\/news\/tencent-hy4-flash-open-weights-hugging-face\">Tencent put Hy4 Flash on Hugging Face<\/a> on 18 November with 224 billion total and 12 billion active parameters. Hy4 is the full-size sibling. It enters a crowded field of very large open-weight models from Chinese labs, including <a href=\"https:\/\/i10x.ai\/news\/moonshot-kimi-k3-1-pro-open-weights\">Moonshot&#8217;s Kimi K3.1 Pro<\/a>.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What&#8217;s in the model<\/h2>\n\n\n\n<ul class=\"wp-block-list\"><li><strong>Architecture:<\/strong> 78 layers, one dense and 77 mixture-of-experts layers, each with 256 routed experts and one shared expert. Every token uses the top 8 routed experts plus the shared one.<\/li><li><strong>Attention:<\/strong> Gated DeepSeek Sparse Attention with IndexCache, which picks the top 2,048 tokens of the context for each step, an approach Tencent says was inspired by DeepSeek and GLM.<\/li><li><strong>Faster output:<\/strong> a built-in 3.3B parameter drafter called DFly2 handles speculative decoding and reuses Hy4&#8217;s own memory cache.<\/li><li><strong>Vision:<\/strong> a 1.23B parameter vision encoder turns images and video frames into tokens for the language model.<\/li><li><strong>Reasoning modes:<\/strong> developers can switch between &#8220;high&#8221; reasoning and &#8220;no_think&#8221; for direct answers.<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">How Tencent says it performs<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">On Tencent&#8217;s internal evaluation set, Hy4 scored 72.53, up from 65.25 for Hy4 preview. In a test of 579 real financial tasks graded by 213 finance professionals, Hy4 averaged 3.04 out of 4, ahead of DeepSeek-V4.1-Flash-1017 at 2.97 and Kimi K3.1 Pro and GLM-5.3 at 2.68 each. Tencent also reports that targeted search upgrades cut the hallucination rate from 15.76% on the preview to 5.45%, and that game development scores rose 19.1%. These are Tencent&#8217;s own numbers, not independent results.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The model card is open about the limits: &#8220;a notable gap remains between it and top proprietary models,&#8221; especially on research-level math, long-horizon coding and advanced visual question answering.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Why it matters<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>For developers<\/strong>, Hy4 is one of the largest multimodal models that can be run in-house, with ready-made recipes for vLLM and SGLang and a GGUF version for testing.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>For companies<\/strong>, open weights mean sensitive documents, screenshots and video can stay on their own infrastructure. The license allows internal use, but a model-as-a-service business with more than $50 million in yearly revenue, or a product with more than 100 million monthly users, needs a separate written license from Tencent.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>For the market<\/strong>, a 770B model with vision and computer use raises the bar for what open-weight releases include, and puts more price pressure on closed API providers.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What&#8217;s still open<\/h2>\n\n\n\n<ul class=\"wp-block-list\"><li>No independent benchmark results for the full Hy4 were available at the time of writing.<\/li><li>The hardware needs (eight top-end GPUs even in FP8) put self-hosting out of reach for most small teams.<\/li><li>How the safety findings in the system card compare with those of other open-weight releases.<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">Sources<\/h2>\n\n\n\n<ul class=\"wp-block-list\"><li>Tencent Hy post on X, 20 November 2026<\/li><li>Hugging Face model cards for tencent\/Hy4, Hy4-FP8 and Hy4-base, and the Tencent Hy License<\/li><li>Tencent Hy research page for Hy4<\/li><li>SGLang cookbook for Hy4; vLLM and Together AI posts on X, 20 November 2026<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83d\udd2d i10x Perspective<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Hy4 shows how quickly Chinese labs are turning flagship models into downloadable assets, now with vision and computer use included. For buyers, the real question is less the benchmark table than the total cost of serving a 770B model versus paying per token. Follow more releases in <a href=\"https:\/\/i10x.ai\/news\/category\/models\">Models<\/a>.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Tencent released the open weights of Hy4 on 20 November 2026, its natively multimodal flagship with 770 billion total and 49 billion active parameters and a 1 million token context. Hy4, an FP8 version and the base model are on Hugging Face, ModelScope and GitCode, with day-one vLLM and SGLang support.<\/p>\n","protected":false},"author":2,"featured_media":9699,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_yoast_wpseo_meta-robots-noindex":"","paper_title":"","paper_url":"","pdf_url":"","doi":"","arxiv_id":"","authors":"","institutions":"","paper_date":"","venue":"","peer_reviewed":false,"code_url":"","data_url":"","funding_coi":"","figure_license":"","paper_license":"","replication":"","study_models":"","study_sample":"","nx_approved":"","footnotes":"","_yoast_wpseo_primary_category":"9","_nx_sec_severity":"","_nx_sec_affected":"","_nx_sec_status":"","_nx_sec_kicker":"","_nx_sec_type":"","_nx_sec_vendor":"","_nx_sec_fixed":"","_nx_sec_action":""},"categories":[9],"tags":[],"companies":[],"products":[],"class_list":["post-9700","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-models"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.7 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Tencent Hy4 Open Weights: 770B Multimodal MoE Model - i10X News<\/title>\n<meta name=\"description\" content=\"Tencent Hy4 is now open-weight: a 770B MoE model with 49B active parameters, vision, computer use and a 1M token context, on Hugging Face under the Tencent Hy License.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Tencent Hy4 Open Weights: 770B Multimodal MoE Model - i10X News\" \/>\n<meta property=\"og:description\" content=\"Tencent Hy4 is now open-weight: a 770B MoE model with 49B active parameters, vision, computer use and a 1M token context, on Hugging Face under the Tencent Hy License.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model\" \/>\n<meta property=\"og:site_name\" content=\"i10X News\" \/>\n<meta property=\"article:published_time\" content=\"2026-11-20T09:43:59+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/11\/tencent-hy4-open-weights-770b-multimodal-model-title.webp\" \/>\n\t<meta property=\"og:image:width\" content=\"1344\" \/>\n\t<meta property=\"og:image:height\" content=\"768\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/webp\" \/>\n<meta name=\"author\" content=\"Christopher Ort\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:image\" content=\"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/i10x-news-og-default-1200x630-1.png\" \/>\n<meta name=\"twitter:creator\" content=\"@i10x_ai\" \/>\n<meta name=\"twitter:site\" content=\"@i10x_ai\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Christopher Ort\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"5 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"NewsArticle\",\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model\"},\"author\":{\"name\":\"Christopher Ort\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#\\\/schema\\\/person\\\/d410bb7ee72febefca2c9c369939219a\"},\"headline\":\"Tencent Hy4 Open Weights: The 770B Multimodal Flagship Is Now Free to Download, With a 1M Token Context\",\"datePublished\":\"2026-11-20T09:43:59+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model\"},\"wordCount\":937,\"publisher\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/tencent-hy4-open-weights-770b-multimodal-model-title.webp\",\"articleSection\":[\"Models\"],\"inLanguage\":\"en-US\",\"dateModified\":\"2026-11-20T09:43:59+00:00\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model\",\"url\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model\",\"name\":\"Tencent Hy4 Open Weights: 770B Multimodal MoE Model - i10X News\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/tencent-hy4-open-weights-770b-multimodal-model-title.webp\",\"datePublished\":\"2026-11-20T09:43:59+00:00\",\"description\":\"Tencent Hy4 is now open-weight: a 770B MoE model with 49B active parameters, vision, computer use and a 1M token context, on Hugging Face under the Tencent Hy License.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model#primaryimage\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/tencent-hy4-open-weights-770b-multimodal-model-title.webp\",\"contentUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/tencent-hy4-open-weights-770b-multimodal-model-title.webp\",\"width\":1344,\"height\":768,\"caption\":\"Image: AI-generated, i10X\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/tencent-hy4-open-weights-770b-multimodal-model#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/i10x.ai\\\/news\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Tencent Hy4 Open Weights: The 770B Multimodal Flagship Is Now Free to Download, With a 1M Token Context\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#website\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\",\"name\":\"i10X News\",\"description\":\"Stay updated with the latest news and announcements from i10X.\",\"publisher\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/i10x.ai\\\/news?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\",\"name\":\"i10X\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/i10x-logo-180.png\",\"contentUrl\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/i10x-logo-180.png\",\"width\":180,\"height\":180,\"caption\":\"i10X\"},\"image\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/x.com\\\/i10x_ai\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/i10x-ai\",\"https:\\\/\\\/instagram.com\\\/i10x.ai\",\"https:\\\/\\\/www.youtube.com\\\/@i10X_ai_official\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#\\\/schema\\\/person\\\/d410bb7ee72febefca2c9c369939219a\",\"name\":\"Christopher Ort\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"contentUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"caption\":\"Christopher Ort\"},\"description\":\"Christopher Ort is the editor of i10X News at i10x.ai. He has worked hands-on with AI tools, models and agent workflows for years and leads i10X\u2019s coverage of AI launches, funding and research. He is the author of \u201cThe Style Penalty\u201d, an SSRN study on how frontier AI models judge AI-written CVs, covered by heise online and t3n.\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/author\\\/christopher-ort\",\"sameAs\":[\"https:\\\/\\\/i10x.ai\\\/blog\\\/author\\\/christopher-ort\"],\"worksFor\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\"},\"alumniOf\":{\"@type\":\"EducationalOrganization\",\"name\":\"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU)\",\"url\":\"https:\\\/\\\/www.fau.eu\"},\"subjectOf\":[{\"@type\":\"CreativeWork\",\"url\":\"https:\\\/\\\/papers.ssrn.com\\\/abstract=7426699\",\"name\":\"The Style Penalty: How AI Resume-Writing Tools Shape Outcomes in Automated Hiring Screens\"},{\"@type\":\"CreativeWork\",\"url\":\"https:\\\/\\\/www.heise.de\\\/news\\\/KI-Chatbots-bewerten-KI-Lebenslaeufe-Claude-findet-sich-besser-als-GPT-11317531.html\",\"name\":\"KI-Chatbots bewerten KI-Lebensl\u00e4ufe: Claude findet sich besser als GPT\"},{\"@type\":\"CreativeWork\",\"url\":\"https:\\\/\\\/t3n.de\\\/news\\\/bewerbung-schreiben-ki-5-fehler-1747190\\\/\",\"name\":\"Bewerbung mit KI schreiben: Diese 5 Fehler solltet ihr laut Studien vermeiden\"}]}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Tencent Hy4 Open Weights: 770B Multimodal MoE Model - i10X News","description":"Tencent Hy4 is now open-weight: a 770B MoE model with 49B active parameters, vision, computer use and a 1M token context, on Hugging Face under the Tencent Hy License.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model","og_locale":"en_US","og_type":"article","og_title":"Tencent Hy4 Open Weights: 770B Multimodal MoE Model - i10X News","og_description":"Tencent Hy4 is now open-weight: a 770B MoE model with 49B active parameters, vision, computer use and a 1M token context, on Hugging Face under the Tencent Hy License.","og_url":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model","og_site_name":"i10X News","article_published_time":"2026-11-20T09:43:59+00:00","og_image":[{"width":1344,"height":768,"url":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/11\/tencent-hy4-open-weights-770b-multimodal-model-title.webp","type":"image\/webp"}],"author":"Christopher Ort","twitter_card":"summary_large_image","twitter_image":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/i10x-news-og-default-1200x630-1.png","twitter_creator":"@i10x_ai","twitter_site":"@i10x_ai","twitter_misc":{"Written by":"Christopher Ort","Est. reading time":"5 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"NewsArticle","@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model#article","isPartOf":{"@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model"},"author":{"name":"Christopher Ort","@id":"https:\/\/i10x.ai\/news\/#\/schema\/person\/d410bb7ee72febefca2c9c369939219a"},"headline":"Tencent Hy4 Open Weights: The 770B Multimodal Flagship Is Now Free to Download, With a 1M Token Context","datePublished":"2026-11-20T09:43:59+00:00","mainEntityOfPage":{"@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model"},"wordCount":937,"publisher":{"@id":"https:\/\/i10x.ai\/news\/#organization"},"image":{"@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/11\/tencent-hy4-open-weights-770b-multimodal-model-title.webp","articleSection":["Models"],"inLanguage":"en-US","dateModified":"2026-11-20T09:43:59+00:00"},{"@type":"WebPage","@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model","url":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model","name":"Tencent Hy4 Open Weights: 770B Multimodal MoE Model - i10X News","isPartOf":{"@id":"https:\/\/i10x.ai\/news\/#website"},"primaryImageOfPage":{"@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model#primaryimage"},"image":{"@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/11\/tencent-hy4-open-weights-770b-multimodal-model-title.webp","datePublished":"2026-11-20T09:43:59+00:00","description":"Tencent Hy4 is now open-weight: a 770B MoE model with 49B active parameters, vision, computer use and a 1M token context, on Hugging Face under the Tencent Hy License.","breadcrumb":{"@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model#primaryimage","url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/11\/tencent-hy4-open-weights-770b-multimodal-model-title.webp","contentUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/11\/tencent-hy4-open-weights-770b-multimodal-model-title.webp","width":1344,"height":768,"caption":"Image: AI-generated, i10X"},{"@type":"BreadcrumbList","@id":"https:\/\/i10xnews.kinsta.cloud\/tencent-hy4-open-weights-770b-multimodal-model#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/i10x.ai\/news"},{"@type":"ListItem","position":2,"name":"Tencent Hy4 Open Weights: The 770B Multimodal Flagship Is Now Free to Download, With a 1M Token Context"}]},{"@type":"WebSite","@id":"https:\/\/i10x.ai\/news\/#website","url":"https:\/\/i10x.ai\/news","name":"i10X News","description":"Stay updated with the latest news and announcements from i10X.","publisher":{"@id":"https:\/\/i10x.ai\/news\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/i10x.ai\/news?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/i10x.ai\/news\/#organization","name":"i10X","url":"https:\/\/i10x.ai\/news\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10x.ai\/news\/#\/schema\/logo\/image\/","url":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/i10x-logo-180.png","contentUrl":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/i10x-logo-180.png","width":180,"height":180,"caption":"i10X"},"image":{"@id":"https:\/\/i10x.ai\/news\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/x.com\/i10x_ai","https:\/\/www.linkedin.com\/company\/i10x-ai","https:\/\/instagram.com\/i10x.ai","https:\/\/www.youtube.com\/@i10X_ai_official"]},{"@type":"Person","@id":"https:\/\/i10x.ai\/news\/#\/schema\/person\/d410bb7ee72febefca2c9c369939219a","name":"Christopher Ort","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","contentUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","caption":"Christopher Ort"},"description":"Christopher Ort is the editor of i10X News at i10x.ai. He has worked hands-on with AI tools, models and agent workflows for years and leads i10X\u2019s coverage of AI launches, funding and research. He is the author of \u201cThe Style Penalty\u201d, an SSRN study on how frontier AI models judge AI-written CVs, covered by heise online and t3n.","url":"https:\/\/i10x.ai\/news\/author\/christopher-ort","sameAs":["https:\/\/i10x.ai\/blog\/author\/christopher-ort"],"worksFor":{"@id":"https:\/\/i10x.ai\/news\/#organization"},"alumniOf":{"@type":"EducationalOrganization","name":"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU)","url":"https:\/\/www.fau.eu"},"subjectOf":[{"@type":"CreativeWork","url":"https:\/\/papers.ssrn.com\/abstract=7426699","name":"The Style Penalty: How AI Resume-Writing Tools Shape Outcomes in Automated Hiring Screens"},{"@type":"CreativeWork","url":"https:\/\/www.heise.de\/news\/KI-Chatbots-bewerten-KI-Lebenslaeufe-Claude-findet-sich-besser-als-GPT-11317531.html","name":"KI-Chatbots bewerten KI-Lebensl\u00e4ufe: Claude findet sich besser als GPT"},{"@type":"CreativeWork","url":"https:\/\/t3n.de\/news\/bewerbung-schreiben-ki-5-fehler-1747190\/","name":"Bewerbung mit KI schreiben: Diese 5 Fehler solltet ihr laut Studien vermeiden"}]}]}},"_links":{"self":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/9700","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/comments?post=9700"}],"version-history":[{"count":0,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/9700\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media\/9699"}],"wp:attachment":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media?parent=9700"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/categories?post=9700"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/tags?post=9700"},{"taxonomy":"company","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/companies?post=9700"},{"taxonomy":"product","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/products?post=9700"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}