[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"nav-categories":3,"article-mistral-medium-3-5-the-128b-multimodal-model-that-actually-fits-on-your-hardware":70},{"data":4},[5,37,57,64],{"name":6,"slug":7,"categories":8},"Productivity","productivity",[9,13,17,21,25,29,33],{"id":10,"title":11,"slug":12},17,"Branding","branding",{"id":14,"title":15,"slug":16},19,"Marketing","marketing",{"id":18,"title":19,"slug":20},20,"Work","work",{"id":22,"title":23,"slug":24},34,"Community","community",{"id":26,"title":27,"slug":28},21,"For newbies","for-newbies",{"id":30,"title":31,"slug":32},24,"Investment","investment",{"id":34,"title":35,"slug":36},22,"Finance","finance",{"name":38,"slug":39,"categories":40},"Tech","tech",[41,45,49,53],{"id":42,"title":43,"slug":44},28,"Technology","technology",{"id":46,"title":47,"slug":48},32,"Artificial Intelligence","artificial-intelligence",{"id":50,"title":51,"slug":52},26,"Security and protection","security-and-protection",{"id":54,"title":55,"slug":56},31,"YouTube Blog","youtube-blog",{"name":58,"slug":59,"categories":60},"News","news",[61],{"id":62,"title":58,"slug":63},18,"quasanews",{"name":65,"slug":66,"categories":67},"Business","business",[68],{"id":69,"title":65,"slug":66},16,{"post":71,"published_news":94,"popular_news":147,"categories":217},{"title":72,"description":73,"meta_title":72,"meta_description":74,"meta_keywords":74,"text":75,"slug":76,"created_at":77,"publish_at":78,"formatted_created_at":79,"category_id":46,"links":80,"view_type":85,"video_url":86,"views":34,"likes":87,"lang":88,"comments_count":87,"category":89},"Mistral Medium 3.5: The 128B Multimodal Model That Actually Fits on Your Hardware","Mistral AI has just released Mistral Medium 3.5 — a dense multimodal model with 128 billion parameters and a massive 256k token context window.","","\u003Cp>\u003Ca href=\"https://quasa.io/projects/mistral-ai\">Mistral AI has just released \u003Cstrong>Mistral Medium 3.5\u003C/strong>\u003C/a> — a dense multimodal model with 128 billion parameters and a massive 256k token context window.\u003C/p>\n\n\u003Cp>\u003Cpicture>\u003Csource srcset=\"https://cdn.quasa.io/photos/0001/image-2026-05-01t095531662.webp\" type=\"image/webp\">\u003Cimg alt=\"Mistral Medium 3.5: The 128B Multimodal Model That Actually Fits on Your Hardware\" class=\"image-align-right\" height=\"402\" src=\"https://cdn.quasa.io/photos/0001/image-2026-05-01t095531662.jpg\" width=\"270\" />\u003C/picture>It’s the latest addition to the company’s growing family of open-weight models, and it strikes a very specific sweet spot in the current LLM landscape.\u003C/p>\n\n\u003Ch4>\u003Cstrong>Performance and Positioning\u003C/strong>\u003C/h4>\n\n\u003Cp>Medium 3.5 clearly outperforms all previous Mistral models. It handles long-context reasoning, multimodal inputs (text + vision), and complex agentic workflows noticeably better than its predecessors.  \u003C/p>\n\n\u003Cp>However, it still doesn’t reach the absolute ceiling set by the very largest open models (think 400B+ parameter beasts). That’s not a bug — it’s the point. Mistral deliberately positioned Medium 3.5 in a weight class where it has almost no direct competition. Most models that match or beat its capabilities are \u003Cstrong>several times larger\u003C/strong>, which makes this 128B model unusually practical for real-world use.\u003C/p>\n\n\u003Chr />\n\u003Ch4>\u003Cstrong>Built for Local Deployment\u003C/strong>\u003C/h4>\n\n\u003Cp>\u003Cpicture>\u003Csource srcset=\"https://cdn.quasa.io/photos/0001/image-2026-05-01t095533429.webp\" type=\"image/webp\">\u003Cimg alt=\"Mistral Medium 3.5: The 128B Multimodal Model That Actually Fits on Your Hardware\" class=\"image-align-left\" height=\"447\" src=\"https://cdn.quasa.io/photos/0001/image-2026-05-01t095533429.jpg\" width=\"300\" />\u003C/picture>This is where Medium 3.5 shines. Because it’s significantly smaller than its performance peers, it becomes a realistic option for local or on-prem deployment on high-end consumer or enterprise hardware.  \u003C/p>\n\n\u003Cp>To make sure it doesn’t crawl like a turtle during inference, Mistral also released a dedicated \u003Cstrong>speculative decoding head\u003C/strong>.\u003C/p>\n\n\u003Cp>This accessory dramatically speeds up generation while keeping quality high — a smart move that turns a theoretically heavy model into something actually usable outside of massive GPU clusters.\u003C/p>\n\n\u003Ch4>\u003Cstrong>Pricing and Cloud Strategy\u003C/strong>\u003C/h4>\n\n\u003Cp>\u003Cstrong>If you’re considering the official API, the numbers are straightforward:  \u003C/strong>\u003C/p>\n\n\u003Cul>\n\t\u003Cli>\u003Cstrong>$1.5 per million input tokens;\u003C/strong>\u003C/li>\n\t\u003Cli>\u003Cstrong>$7.5 per million output tokens.\u003C/strong>\u003C/li>\n\u003C/ul>\n\n\u003Cp>At those prices, there’s very little incentive to run Medium 3.5 in the cloud. The model was clearly designed to be \u003Cstrong>self-hosted\u003C/strong>. The open weights are already available on Hugging Face, and the economics strongly favor downloading and running it yourself.\u003C/p>\n\n\u003Chr />\n\u003Ch4>\u003Cstrong>Licensing\u003C/strong>\u003C/h4>\n\n\u003Cp>\u003Cpicture>\u003Csource srcset=\"https://cdn.quasa.io/photos/0001/image-2026-05-01t095535197.webp\" type=\"image/webp\">\u003Cimg alt=\"Mistral Medium 3.5: The 128B Multimodal Model That Actually Fits on Your Hardware\" class=\"image-align-right\" height=\"417\" src=\"https://cdn.quasa.io/photos/0001/image-2026-05-01t095535197.jpg\" width=\"280\" />\u003C/picture>The license is genuinely open for most users and smaller companies. However, \u003Ca href=\"https://quasa.io/projects/mistral-ai\">Mistral\u003C/a> added a commercial clause: any organization with \u003Cstrong>more than $20 million in monthly revenue\u003C/strong> must purchase an enterprise license.\u003C/p>\n\n\u003Cp>It’s a clean, predictable rule that protects Mistral’s business model without scaring away the broader developer community.\u003C/p>\n\n\u003Cp>Also read:\u003C/p>\n\n\u003Cul>\n\t\u003Cli>\u003Ca href=\"https://quasa.io/media/375-billion-underfunded-atomico-shows-why-europe-s-vc-remains-fragmented-and-government-heavy\">$375 Billion Underfunded: Atomico Shows Why Europe’s VC Remains Fragmented and Government-Heavy\u003C/a>\u003C/li>\n\t\u003Cli>\u003Ca href=\"https://quasa.io/media/mrbeast-s-bold-leap-into-fintech-acquiring-step-amidst-business-challenges\">MrBeast's Bold Leap into Fintech: Acquiring Step Amidst Business Challenges\u003C/a>\u003C/li>\n\t\u003Cli>\u003Ca href=\"https://quasa.io/media/ai-hype-vs-reality-deloitte-s-tech-trends-2026-exposes-the-gap-between-talk-and-deployment\">AI Hype vs. Reality: Deloitte's Tech Trends 2026 Exposes the Gap Between Talk and Deployment\u003C/a>\u003C/li>\n\u003C/ul>\n\n\u003Chr />\n\u003Ch4>\u003Cstrong>Links\u003C/strong>\u003C/h4>\n\n\u003Cul>\n\t\u003Cli>\u003Ca href=\"https://huggingface.co/collections/mistralai/mistral-medium-35\">\u003Cstrong>Model weights & collection\u003C/strong>: https://huggingface.co/collections/mistralai/mistral-medium-35 \u003C/a> \u003C/li>\n\t\u003Cli>\u003Ca href=\"https://mistral.ai/news/vibe-remote-agents-mistral-medium-3-5\">\u003Cstrong>Official announcement\u003C/strong>: https://mistral.ai/news/vibe-remote-agents-mistral-medium-3-5\u003C/a>\u003C/li>\n\u003C/ul>\n\n\u003Chr />\n\u003Ch4>\u003Cstrong>The Bottom Line\u003C/strong>\u003C/h4>\n\n\u003Cp>Mistral Medium 3.5 isn’t trying to be the biggest model on the leaderboard. It’s trying to be the \u003Cstrong>most usable high-performance model\u003C/strong> in its size range — and it largely succeeds.\u003C/p>\n\n\u003Cp>\u003Cpicture>\u003Csource srcset=\"https://cdn.quasa.io/photos/0001/photo_2026-05-01_09-49-44.webp\" type=\"image/webp\">\u003Cimg class=\"image-align-center\"  alt=\"Mistral Medium 3.5: The 128B Multimodal Model That Actually Fits on Your Hardware\" height=\"293\" src=\"https://cdn.quasa.io/photos/0001/photo_2026-05-01_09-49-44.jpg\" width=\"400\" />\u003C/picture>\u003C/p>\n\n\u003Cp>For developers, researchers, and companies that want frontier-level capabilities without needing a supercomputer or paying premium cloud rates, this release is genuinely exciting. It continues Mistral’s pattern of shipping practical, no-nonsense models that close the gap between “impressive benchmark” and “actually runnable on real hardware.”\u003C/p>\n\n\u003Cp>If you’re into local LLMs or building production agents that need long context and multimodal understanding, Medium 3.5 deserves a serious look. The era of “small but mighty” is getting stronger — and Mistral just made it a lot more interesting.\u003C/p>","mistral-medium-3-5-the-128b-multimodal-model-that-actually-fits-on-your-hardware","2026-05-01T08:03:24.000000Z","2026-05-03T09:50:00.000000Z","03.05.2026",{"image":81,"image_webp":82,"thumb":83,"thumb_webp":84},"https://cdn.quasa.io/images/news/aro7Vby9EsnUkMSpyy0aBzdNELBrRmVlYCU7fikM.jpg","https://cdn.quasa.io/images/news/aro7Vby9EsnUkMSpyy0aBzdNELBrRmVlYCU7fikM.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/aro7Vby9EsnUkMSpyy0aBzdNELBrRmVlYCU7fikM.jpg","https://cdn.quasa.io/thumbs/news-thumb/images/news/aro7Vby9EsnUkMSpyy0aBzdNELBrRmVlYCU7fikM.webp","small",null,0,"en",{"id":46,"title":47,"slug":48,"meta_title":90,"meta_description":91,"meta_keywords":91,"deleted_at":86,"created_at":92,"updated_at":93,"lang":88},"Artificial Intelligence | AI Breakthroughs, Agents & Tools | QUASA","Artificial Intelligence, ai, ml, machine learning, chatgpt, future","2024-09-22T08:08:27.000000Z","2026-04-22T14:56:34.000000Z",[95,108,120,133,144],{"title":96,"description":97,"slug":98,"created_at":99,"publish_at":99,"formatted_created_at":79,"category":100,"links":101,"view_type":85,"video_url":86,"views":106,"likes":87,"lang":88,"comments_count":87,"is_pinned":107},"To Make AI Work Well – Give It a Mortgage","So yes, the punchline is serious: if you want AI to run your company like a human who actually cares, maybe the first thing you should do is hand it a mortgage.","to-make-ai-work-well-give-it-a-mortgage","2026-05-03T13:21:05.000000Z",{"title":47,"slug":48},{"image":102,"image_webp":103,"thumb":104,"thumb_webp":105},"https://cdn.quasa.io/images/news/t9pvGcKjslgleePmCo7Y6r4GcjOVCcV6SxCFOCaY.jpg","https://cdn.quasa.io/images/news/t9pvGcKjslgleePmCo7Y6r4GcjOVCcV6SxCFOCaY.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/t9pvGcKjslgleePmCo7Y6r4GcjOVCcV6SxCFOCaY.jpg","https://cdn.quasa.io/thumbs/news-thumb/images/news/t9pvGcKjslgleePmCo7Y6r4GcjOVCcV6SxCFOCaY.webp",1,false,{"title":109,"description":110,"slug":111,"created_at":112,"publish_at":112,"formatted_created_at":79,"category":113,"links":114,"view_type":85,"video_url":86,"views":119,"likes":87,"lang":88,"comments_count":87,"is_pinned":107},"Scambodia vs. Scammy Korea: Two Flavors of National Cyber-Scam Empires","In a deeply reported piece from April 2026, WSJ lays out how Cambodia has quietly built one of the world’s most efficient criminal enterprises.","scambodia-vs-scammy-korea-two-flavors-of-national-cyber-scam-empires","2026-05-03T12:17:49.000000Z",{"title":58,"slug":63},{"image":115,"image_webp":116,"thumb":117,"thumb_webp":118},"https://cdn.quasa.io/images/news/Z0PjcPumMEqK7RQoNFaNwnnlVuvyrXCqc2UsAqHS.jpg","https://cdn.quasa.io/images/news/Z0PjcPumMEqK7RQoNFaNwnnlVuvyrXCqc2UsAqHS.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/Z0PjcPumMEqK7RQoNFaNwnnlVuvyrXCqc2UsAqHS.jpg","https://cdn.quasa.io/thumbs/news-thumb/images/news/Z0PjcPumMEqK7RQoNFaNwnnlVuvyrXCqc2UsAqHS.webp",7,{"title":121,"description":122,"slug":123,"created_at":124,"publish_at":125,"formatted_created_at":79,"category":126,"links":127,"view_type":85,"video_url":86,"views":132,"likes":87,"lang":88,"comments_count":87,"is_pinned":107},"Alibaba Just Launched Happy Horse 1.0 – A Serious Budget Challenger to Seedance 2","Alibaba has quietly entered the crowded AI video generation space with Happy Horse 1.0, now available in limited beta.","alibaba-just-launched-happy-horse-1-0-a-serious-budget-challenger-to-seedance-2","2026-05-01T08:57:17.000000Z","2026-05-03T11:12:00.000000Z",{"title":43,"slug":44},{"image":128,"image_webp":129,"thumb":130,"thumb_webp":131},"https://cdn.quasa.io/images/news/KgBTVgWK0hCKJOaKRMtaxvrSgpp0zLEMV3ka9A77.jpg","https://cdn.quasa.io/images/news/KgBTVgWK0hCKJOaKRMtaxvrSgpp0zLEMV3ka9A77.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/KgBTVgWK0hCKJOaKRMtaxvrSgpp0zLEMV3ka9A77.jpg","https://cdn.quasa.io/thumbs/news-thumb/images/news/KgBTVgWK0hCKJOaKRMtaxvrSgpp0zLEMV3ka9A77.webp",15,{"title":134,"description":135,"slug":136,"created_at":137,"publish_at":137,"formatted_created_at":79,"category":138,"links":139,"view_type":85,"video_url":86,"views":69,"likes":87,"lang":88,"comments_count":87,"is_pinned":107},"The Creator of AlphaGo Just Raised $1.1 Billion on One Radical Thesis — And It Could Redefine the Entire Future of AI","Old-school AI watchers still get chills remembering 2017. That was the year DeepMind dropped AlphaZero — the spiritual successor to the legendary AlphaGo that had already shocked the world by beating the best human Go players. But AlphaZero was something far more audacious.","the-creator-of-alphago-just-raised-1-1-billion-on-one-radical-thesis-and-it-could-redefine-the-entire-future-of-ai","2026-05-03T10:57:04.000000Z",{"title":43,"slug":44},{"image":140,"image_webp":141,"thumb":142,"thumb_webp":143},"https://cdn.quasa.io/images/news/L5URa8knNUENc2LFD3l7UdyHAXW5TbvcVEPryEcl.jpg","https://cdn.quasa.io/images/news/L5URa8knNUENc2LFD3l7UdyHAXW5TbvcVEPryEcl.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/L5URa8knNUENc2LFD3l7UdyHAXW5TbvcVEPryEcl.jpg","https://cdn.quasa.io/thumbs/news-thumb/images/news/L5URa8knNUENc2LFD3l7UdyHAXW5TbvcVEPryEcl.webp",{"title":72,"description":73,"slug":76,"created_at":77,"publish_at":78,"formatted_created_at":79,"category":145,"links":146,"view_type":85,"video_url":86,"views":34,"likes":87,"lang":88,"comments_count":87,"is_pinned":107},{"title":47,"slug":48},{"image":81,"image_webp":82,"thumb":83,"thumb_webp":84},[148,161,177,189,203],{"title":149,"description":150,"slug":151,"created_at":152,"publish_at":153,"formatted_created_at":154,"category":155,"links":156,"view_type":85,"video_url":86,"views":159,"likes":160,"lang":88,"comments_count":87,"is_pinned":107},"The Anatomy of an Entrepreneur","Entrepreneur is a French word that means an enterpriser. Enterprisers are people who undertake a business or enterprise with the chance of earning profits or suffering from loss.","the-anatomy-of-an-entrepreneur","2021-08-04T15:18:21.000000Z","2025-12-14T06:09:00.000000Z","14.12.2025",{"title":65,"slug":66},{"image":157,"image_webp":86,"thumb":158,"thumb_webp":158},"https://cdn.quasa.io/images/news/mVsXPTMuHZuI7UXCsENgL1Qwp1uSOf7Rz3uVPMfm.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/mVsXPTMuHZuI7UXCsENgL1Qwp1uSOf7Rz3uVPMfm.webp",72623,2,{"title":162,"description":163,"slug":164,"created_at":165,"publish_at":166,"formatted_created_at":167,"category":168,"links":169,"view_type":174,"video_url":86,"views":175,"likes":176,"lang":88,"comments_count":87,"is_pinned":107},"Advertising on QUASA","QUASA MEDIA is read by more than 400 thousand people a month. We offer to place your article, add a link or order the writing of an article for publication.","advertising-on-quasa","2022-07-06T07:33:02.000000Z","2025-12-15T17:33:02.000000Z","15.12.2025",{"title":58,"slug":63},{"image":170,"image_webp":171,"thumb":172,"thumb_webp":173},"https://cdn.quasa.io/images/news/45SvmdsTQbiyc3nxgbyHY1mpVbisYyub2BCHjqBL.jpg","https://cdn.quasa.io/images/news/45SvmdsTQbiyc3nxgbyHY1mpVbisYyub2BCHjqBL.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/45SvmdsTQbiyc3nxgbyHY1mpVbisYyub2BCHjqBL.jpg","https://cdn.quasa.io/thumbs/news-thumb/images/news/45SvmdsTQbiyc3nxgbyHY1mpVbisYyub2BCHjqBL.webp","large",72414,4,{"title":178,"description":179,"slug":180,"created_at":181,"publish_at":182,"formatted_created_at":183,"category":184,"links":185,"view_type":85,"video_url":86,"views":188,"likes":176,"lang":88,"comments_count":87,"is_pinned":107},"What is a Startup?","A startup is not a new company, not a tech company, nor a new tech company. You can be a new tech company, if your goal is not to grow high and fast; then, you are not a startup. ","what-is-a-startup","2021-08-04T12:05:17.000000Z","2025-12-17T13:02:00.000000Z","17.12.2025",{"title":65,"slug":66},{"image":186,"image_webp":86,"thumb":187,"thumb_webp":187},"https://cdn.quasa.io/images/news/EOsQhSW3VXyG7a6NPdE1oZd00xfJXe3bjY5aJGb7.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/EOsQhSW3VXyG7a6NPdE1oZd00xfJXe3bjY5aJGb7.webp",69973,{"title":190,"description":191,"slug":192,"created_at":193,"publish_at":194,"formatted_created_at":195,"category":196,"links":197,"view_type":85,"video_url":86,"views":202,"likes":160,"lang":88,"comments_count":106,"is_pinned":107},"Top 5 Tips to Make More Money as a Content Creator","Content creators are one of the most desired job titles right now. Who wouldn’t want to earn a living online?","top-5-tips-to-make-more-money-as-a-content-creator","2022-01-17T17:31:51.000000Z","2026-01-17T11:30:00.000000Z","17.01.2026",{"title":19,"slug":20},{"image":198,"image_webp":199,"thumb":200,"thumb_webp":201},"https://cdn.quasa.io/images/news/gP8kiumBPpJmQv6SMieXiX1tDetx43VwFfO1P4Ca.jpg","https://cdn.quasa.io/images/news/gP8kiumBPpJmQv6SMieXiX1tDetx43VwFfO1P4Ca.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/gP8kiumBPpJmQv6SMieXiX1tDetx43VwFfO1P4Ca.jpg","https://cdn.quasa.io/thumbs/news-thumb/images/news/gP8kiumBPpJmQv6SMieXiX1tDetx43VwFfO1P4Ca.webp",43831,{"title":204,"description":205,"slug":206,"created_at":207,"publish_at":208,"formatted_created_at":209,"category":210,"links":211,"view_type":174,"video_url":86,"views":216,"likes":160,"lang":88,"comments_count":87,"is_pinned":107},"8 Logo Design Tips for Small Businesses","Your logo tells the story of your business and the values you stand for.","8-logo-design-tips-for-small-businesses","2021-12-04T21:59:52.000000Z","2025-05-05T03:30:00.000000Z","05.05.2025",{"title":15,"slug":16},{"image":212,"image_webp":213,"thumb":214,"thumb_webp":215},"https://cdn.quasa.io/images/news/Wbx2NtS1CnTupgoQbpFMGspJ5jm4uob2hDOq33r0.jpg","https://cdn.quasa.io/images/news/Wbx2NtS1CnTupgoQbpFMGspJ5jm4uob2hDOq33r0.webp","https://cdn.quasa.io/thumbs/news-thumb/images/news/Wbx2NtS1CnTupgoQbpFMGspJ5jm4uob2hDOq33r0.jpg","https://cdn.quasa.io/thumbs/news-thumb/images/news/Wbx2NtS1CnTupgoQbpFMGspJ5jm4uob2hDOq33r0.webp",42819,[218,219,220,221,222,223,224,225,226,227,228,229,230],{"title":23,"slug":24},{"title":47,"slug":48},{"title":55,"slug":56},{"title":43,"slug":44},{"title":51,"slug":52},{"title":31,"slug":32},{"title":35,"slug":36},{"title":27,"slug":28},{"title":19,"slug":20},{"title":15,"slug":16},{"title":58,"slug":63},{"title":11,"slug":12},{"title":65,"slug":66}]