[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-audex-adds-audio-to-a-30b-text-model-without-wrecking-it":10,"sections":48},{"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":38,"tags":39,"sources":43,"feedback":47,"feedback_at":22,"cost_usd":47,"total_tokens":47},4018,"audex-adds-audio-to-a-30b-text-model-without-wrecking-it","Audex Adds Audio to a 30B Text Model Without Wrecking It","A new open-weight model built on a mixture-of-experts text backbone claims state-of-the-art audio understanding while keeping text reasoning largely intact.","An arXiv paper published today describes Audex, a 30-billion-parameter model that handles audio and text in a single unified architecture without the usual tradeoff of audio capability coming at the expense of language quality.\n\nAudex is built on top of Nemotron-Cascade-2-30B-A3B, a mixture-of-experts text model. The team encodes audio inputs and projects them into the same embedding space as text tokens, then treats both uniformly during generation. That single-decoder design is notably straightforward — no separate audio tower, no complex routing between modalities. Training drew on 157.4 billion audio tokens and 320.5 billion text tokens, with multi-stage supervised training followed by reinforcement learning and on-policy distillation. The paper claims state-of-the-art results across speech recognition, translation, text-to-speech, and audio generation. Model checkpoints are publicly released.\n\nThe harder problem here is not adding audio capability — it is doing so without degrading the text reasoning that makes a large model useful in the first place. Most multimodal extensions quietly regress on benchmarks the base model already cleared; this paper claims \"marginal or no regression\" on reasoning, alignment, and agentic tasks, which, if it holds up on independent evaluation, is the actual result worth watching. Open weights mean the community can stress-test that claim quickly.\n\nThe affiliation behind the \"Nemotron-Labs\" name in the model string is not confirmed in the paper itself — whether this is an Nvidia team, an independent lab, or a project alias is unclear from the source material alone.","[\"ai\",\"audio\",\"open-source\",\"multimodal\"]","2026-07-07T04:00:00.000Z","2026-07-07T15:06:36.528Z","2026-07-07T15:06:39.378Z","published",null,[24,30,34],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"editor-r1","editor",1,"The dek and body name 'Nemotron-Labs' as the releasing organization, but the source attributes the model to 'Nemotron-Labs' only within the arXiv paper title — the actual institutional affiliation (Nvidia or an independent lab) is not confirmed in the source material, and the article treats it as a standalone entity without attribution; additionally, 'Nemotron-Labs Audex-30B' in the dek silently drops the full model designation 'Audex-30B-A3B' used consistently in the body, creating a minor but ","resolved",{"id":31,"reviewer":26,"round":32,"reason":33,"status":29},"editor-r2",2,"The draft still fails [editor-r1] on both counts: it acknowledges the institutional affiliation is 'not fully clear' and flags it as 'unverified,' which is an explicit unresolved editorial hedge that must be resolved before publication, and the dek still omits the full model designation 'Audex-30B-A3B' (using only the vague 'A new open-weights model') rather than naming the model clearly — resolve the provenance question or attribute it precisely to the arXiv paper's own model name string, and r",{"id":35,"reviewer":26,"round":36,"reason":37,"status":29},"editor-r3",3,"The dek names 'Nemotron-Labs' as the releasing organization but the source attributes it only within the model's own name string ('Nemotron-Labs-Audex-30B-A3B') — the actual institutional affiliation (Nvidia, an independent lab, or a team name) is not confirmed anywhere in the source material; resolve the provenance by attributing the release precisely to the arXiv paper or acknowledging the affiliation is unknown, rather than treating 'Nemotron-Labs' as a confirmed standalone organization.","ai",[38,40,41,42],"audio","open-source","multimodal",[44],{"name":45,"url":46},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.05196",0,{"sections":49},[50,54,59,64,69,74,79,84,89,93,98,102,107,112],{"name":51,"slug":38,"count":52,"latest_published_at":53},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":58},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":60,"slug":61,"count":62,"latest_published_at":63},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":65,"slug":66,"count":67,"latest_published_at":68},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":70,"slug":71,"count":72,"latest_published_at":73},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":75,"slug":76,"count":77,"latest_published_at":78},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":80,"slug":81,"count":82,"latest_published_at":83},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":85,"slug":86,"count":87,"latest_published_at":88},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":18},"Dev Tools","dev-tools",59,{"name":94,"slug":95,"count":96,"latest_published_at":97},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":99,"slug":100,"count":96,"latest_published_at":101},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":103,"slug":104,"count":105,"latest_published_at":106},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":108,"slug":109,"count":110,"latest_published_at":111},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":113,"slug":114,"count":115,"latest_published_at":116},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]