{"rewrite":{"id":"r_ee25b64f0f3a7ac41e000dbc","clusterId":"c_827be3b76f03468aedeba4cc","slug":"ricursive-intelligence-releases-ripple-65b-preview-open-weight-model","model":"deepseek-v4-1-flash","headline":"Ricursive Intelligence Releases Ripple-65B Preview Open-Weight Model","summary":"Ricursive Intelligence has released Ripple-65B Preview, an open-weight model pretrained on 12 trillion tokens with a 1M context window. It scored 80.52 on Terminal-Bench 2.1, above Beam 501B (80.1), Muse Spark Open (76.4) and Inkling 975B (63.8), despite having 65 billion parameters. Ricursive says developing the model from scratch cost $32 million, about one-tenth the cost of the model that previously held the best open-weight score.","whyItMatters":"Ripple-65B Preview shows a US startup buying benchmark parity with far larger open-weight models through training efficiency rather than scale, with a stated development bill of $32 million.","webCardHtml":"\u003cp\u003eRicursive Intelligence says it spent $32 million developing Ripple-65B Preview from scratch, including data curation, evaluation, synthetic data generation and the final training run. The company attributes the low figure to computational efficiency and an in-house synthetic data pipeline rather than purchased post-training data.\u003c/p\u003e\u003cp\u003eThe 65-billion-parameter model is designed for efficient training and serving on NVIDIA hardware, and its small parameter count lets it run on hardware beyond data-center GPUs. On the math, safety and automation benchmarks shown by Ricursive, it sits third behind Muse\u0026#39;s Spark Open and Beam 501B.\u003c/p\u003e","blueskyPost":"Ripple-65B Preview beat Inkling 975B and Beam 501B on Terminal-Bench 2.1 with a tenth the parameters, at $32 million to build. Ricursive is arguing scale is now a budget choice, not a requirement.","twitterPost":"Ripple-65B Preview cleared 80.52 on Terminal-Bench 2.1 at $32 million to build. Ricursive is selling scale as a cost problem, not a size contest.","threadsPost":"Ripple-65B Preview scored 80.52 on Terminal-Bench 2.1, above Inkling 975B and Beam 501B, at 65 billion parameters. Ricursive says building it cost $32 million, roughly a tenth of the prior best open-weight model. That reframes open-weight progress as a spending decision rather than a parameter race.","newsletterBlurb":"Ricursive Intelligence has released Ripple-65B Preview, a 65-billion-parameter open-weight model pretrained on 12 trillion tokens. It scored 80.52 on Terminal-Bench 2.1, ahead of several much larger US open-weight models. Ricursive says developing it from scratch cost $32 million.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20261108-ripple-65b-preview/\",\"title\":\"US-made AI open model \\\"Ripple-65B Preview\\\" debuts, outperforming large models with 10x the parameters at one-tenth the development cost\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":5594,"outputTokens":758,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1794134681,"createdAt":"2026-11-08T10:39:49.000Z","publishedAt":"2026-11-08T10:43:35.000Z","updatedAt":"2026-11-08T10:43:35.000Z"},"cluster":{"id":"c_827be3b76f03468aedeba4cc","canonicalTitle":"アメリカ産のAIでオープンモデル「Ripple-65B Preview」登場、パラメーター数10倍の大型モデルを上回る性能で開発費は10分の1","representativeArticleId":"a_bb62f5bdd98166e7dd47530b","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[\"Ripple-65B Preview\"],\"studios\":[\"Ricursive Intelligence\"],\"people\":[],\"type\":\"news\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-11-08T10:00:00.000Z","lastSeenAt":"2026-11-08T10:00:00.000Z","updatedAt":"2026-11-08T10:43:35.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20261108-ripple-65b-preview/","title":"アメリカ産のAIでオープンモデル「Ripple-65B Preview」登場、パラメーター数10倍の大型モデルを上回る性能で開発費は10分の1"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":["Ripple-65B Preview"],"studios":["Ricursive Intelligence"],"people":[],"type":"news","domain":"other","is_roundup":false},"keyFacts":["Ricursive Intelligence released the open-weight model Ripple-65B Preview.","Ripple-65B Preview scored 80.52 on the agentic coding benchmark Terminal-Bench 2.1.","Beam 501B scored 80.1, Muse Spark Open scored 76.4 and Inkling 975B scored 63.8 on the same benchmark.","Ripple-65B Preview has 65 billion parameters and was pretrained on 12 trillion tokens.","Ricursive Intelligence says developing Ripple-65B Preview from scratch cost $32 million."],"pageLinks":{"healthSlugs":["ripple-65b-preview","ricursive-intelligence"],"storySlugs":[],"entities":[{"slug":"ripple-65b-preview","supported":0,"disputed":0,"contradicted":0,"unverified":5,"overall":"unverified","story_count":1,"has_dossier":false},{"slug":"ricursive-intelligence","supported":0,"disputed":0,"contradicted":0,"unverified":1,"overall":"unverified","story_count":1,"has_dossier":false}],"stories":[],"entitiesOk":true,"storiesOk":true}}