[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-vtbench-finds-ai-image-tokenizers-still-trail-continuous-vaes":10,"sections":34},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},10091,"vtbench-finds-ai-image-tokenizers-still-trail-continuous-vaes","VTBench Finds AI Image Tokenizers Still Trail Continuous VAEs","A new benchmark isolates visual tokenizers from the models they feed, and the discrete versions used in most autoregressive image generators come up short.","A new benchmark says the discrete tokenizers powering most autoregressive image generators still can't match plain old VAEs at the basics: reconstructing an image faithfully.\n\nResearchers built VTBench, a benchmark that tests visual tokenizers (the component that turns pixels into discrete tokens before an autoregressive model generates an image) on three tasks: image reconstruction, detail preservation, and text preservation. They ran today's state-of-the-art tokenizers through a battery of metrics and found that continuous variational autoencoders, or VAEs, preserve spatial structure and semantic detail far better than discrete tokenizers do. The discrete versions tend to blur textures, distort objects, and mangle text in reconstructed images. The team also examined GPT-4o's image generation feature, discussing evidence that it may work in an autoregressive way, and released the benchmark and code publicly.\n\nVisual tokenizer quality sets the ceiling on how good an autoregressive image model can get, and this gap helps explain why AR generators still mangle text and fine detail compared with diffusion models built on VAEs. It is also a quiet rebuttal to the assumption that tokenization is a solved problem in autoregressive image generation: the researchers found no open-source tokenizer that closes the gap, and they are calling on the community to build better ones.\n\nUntil someone ships a discrete tokenizer that can actually read a street sign, autoregressive image generation will keep looking better in the abstract than on the page.","[\"ai\",\"image-generation\",\"benchmarks\",\"visual-tokenizers\"]","2026-10-05T04:00:00.000Z","2026-10-05T22:36:01.455Z","2026-10-05T22:36:07.275Z","published",null,[],"ai",[24,26,27,28],"image-generation","benchmarks","visual-tokenizers",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.13439",0,{"sections":35},[36,39,43,48,53,58,62,67,71,76,81,86,91,96],{"name":37,"slug":24,"count":38,"latest_published_at":18},"AI",6314,{"name":40,"slug":41,"count":42,"latest_published_at":18},"Security","security",871,{"name":44,"slug":45,"count":46,"latest_published_at":47},"Policy","policy",444,"2026-10-03T15:02:01.000Z",{"name":49,"slug":50,"count":51,"latest_published_at":52},"Deals","deals",325,"2026-10-04T13:00:00.000Z",{"name":54,"slug":55,"count":56,"latest_published_at":57},"Hardware","hardware",204,"2026-10-03T14:50:50.000Z",{"name":59,"slug":60,"count":61,"latest_published_at":18},"Science","science",179,{"name":63,"slug":64,"count":65,"latest_published_at":66},"Consumer Tech","consumer-tech",158,"2026-10-03T03:21:12.000Z",{"name":68,"slug":69,"count":70,"latest_published_at":18},"Dev Tools","dev-tools",98,{"name":72,"slug":73,"count":74,"latest_published_at":75},"Software","software",97,"2026-10-04T10:00:00.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Startups","startups",92,"2026-10-04T14:36:25.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Gaming","gaming",53,"2026-10-02T02:50:39.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"General","general",51,"2026-10-05T02:35:01.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"Reviews","reviews",32,"2026-10-02T18:00:00.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"How-To","how-to",7,"2026-10-01T09:00:00.000Z"]