[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-writers-palmyra-x6-tops-agent-benchmarks-with-tiny-training-set":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},5469,"writers-palmyra-x6-tops-agent-benchmarks-with-tiny-training-set","Writer's Palmyra x6 Tops Agent Benchmarks With Tiny Training Set","Writer trained its new Palmyra x6 model on just 626 examples and still claims the top score on a leading tool-use benchmark.","Writer's newest model just claimed the top score on a widely watched benchmark for AI agents that use tools - and it did that with a training set you could read in an afternoon.\n\nPalmyra x6 is a Mixture-of-Experts model that Writer post-trained using a technique it calls Anchored Supervised Fine-Tuning. The recipe is small on purpose: 626 verified, synthetic tool-use examples, a single pass through the data, a low learning rate, and a KL anchor that keeps the model from drifting too far from its frozen base. Writer says the result outperforms its previous default Writer Agent model by a wide margin, scores 0.785 on BFCL Core (the highest in its comparison group), and posts the best average across six public benchmarks. The company also reports it holds up well on internal bias and safety evaluations against the same comparators.\n\nThat's notable because most model post-training stories lean on scale: more data, more compute, more RLHF rounds. Writer is making the opposite bet - that a small, tightly controlled fine-tune, anchored to the base model's original behavior, can beat brute force for a narrow, high-stakes skill like calling the right tool with the right arguments. For enterprise buyers who care as much about predictability as raw capability, that's arguably a better pitch than another leaderboard chart.\n\nStill, this is Writer grading its own homework: the headline numbers come from Writer's own report, on benchmarks it chose, against a comparator list it picked. Every vendor selling agent tooling currently claims to be the best at exactly this. Treat the 0.785 as a data point, not a verdict.","[\"ai\",\"llm-benchmarks\",\"enterprise-ai\",\"fine-tuning\"]","2026-08-18T04:00:00.000Z","2026-08-18T21:33:09.167Z","2026-08-18T21:33:21.060Z","published",null,[],"ai",[24,26,27,28],"llm-benchmarks","enterprise-ai","fine-tuning",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2608.16620",0,{"sections":35},[36,40,44,49,54,59,64,69,74,78,83,88,93,98],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",3293,"2026-08-20T04:00:00.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":39},"Security","security",435,{"name":45,"slug":46,"count":47,"latest_published_at":48},"Policy","policy",210,"2026-08-19T09:32:27.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":58},"Hardware","hardware",140,"2026-08-19T18:25:42.000Z",{"name":60,"slug":61,"count":62,"latest_published_at":63},"Consumer Tech","consumer-tech",95,"2026-08-18T16:05:00.000Z",{"name":65,"slug":66,"count":67,"latest_published_at":68},"Science","science",90,"2026-08-19T18:41:02.000Z",{"name":70,"slug":71,"count":72,"latest_published_at":73},"Software","software",73,"2026-08-18T07:51:50.000Z",{"name":75,"slug":76,"count":77,"latest_published_at":18},"Dev Tools","dev-tools",69,{"name":79,"slug":80,"count":81,"latest_published_at":82},"Startups","startups",47,"2026-08-19T19:13:46.000Z",{"name":84,"slug":85,"count":86,"latest_published_at":87},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":89,"slug":90,"count":91,"latest_published_at":92},"General","general",33,"2026-08-18T22:18:13.000Z",{"name":94,"slug":95,"count":96,"latest_published_at":97},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":99,"slug":100,"count":101,"latest_published_at":102},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]