{"id":"ripple-65b-preview","slug":"ripple-65b-preview","name":"Ripple-65B Preview","orgId":"ricursive","family":"ripple-65b-preview","releaseDate":"2026-11-04","availability":"ga","openWeights":true,"license":"Apache-2.0 (not gated)","paramsTotal":65000000000,"paramsSource":"epoch","contextWindow":1048576,"knowledgeCutoff":null,"modelCardUrl":"https://www.ricursive.com/ripple","hfRepo":"ricursiveAI/Ripple-65B-Preview","modalities":["text"],"aliases":["Ripple-65B Preview","ripple-65b-preview","ripple 65b preview","ripple-65b","ricursiveAI/Ripple-65B-Preview","ricursiveAI/Ripple-65B-Base-Preview","Ripple-65B-Base-Preview","Ripple-65B-Preview-NVFP4","Ripple-65B-Preview-DFlash2","Ricursive: Ripple 65B Preview","ricursiveai/ripple-65b-preview"],"aliasNote":null,"origin":"tracked","tracked":{"trackedSince":"2026-11-04","announced":"2026-11-04","blurb":"Ricursive Intelligence's dense 65B open-weight language model for agentic coding, pretrained from scratch on 12T tokens. A general-purpose variant of the model behind Ricursive's chip-design platform. Released as an instruct model with a base checkpoint, under Apache-2.0.","api":"live","openWeights":"yes","params":{"text":"64.78B dense (64,782,689,360)","source":"Hugging Face safetensors metadata"},"facts":[{"label":"Architecture","value":"Dense transformer, 80 layers, d_model 5120 (Gemma 3N-like, per the model card)"},{"label":"Pretraining","value":"From scratch on 12T tokens at 8K sequence length, starting from an Olmo 3 baseline recipe"},{"label":"Context","value":"Up to 1M tokens"},{"label":"License","value":"Apache-2.0 (not gated)"},{"label":"Checkpoints","value":"Instruct (BF16), Base (pretrain-only), NVFP4 and DFlash2 variants on Hugging Face; one page covers all"},{"label":"Hosted access","value":"Together AI, GMI Cloud and OpenRouter at launch; Fireworks announced for later the same day"}],"claims":["Lab claim: 80.52% on Terminal-Bench 2.1, 79.1% on SWE-Bench Pro v2.0-Hard and 44.3% on DeepSWE v1.1.","Lab claim: about 250 tokens/s on 2x B200 at batch size 8, and small enough to run on a MacBook or NVIDIA DGX Spark.","Lab claim: a 16x compute-efficiency gain on code relative to the Olmo 3 scaling ladder, with a total compute budget of about $32M."],"sources":[{"kind":"lab","title":"Introducing Ripple-65B Preview","url":"https://www.ricursive.com/ripple","date":"2026-11-04"},{"kind":"lab","title":"Ricursive Intelligence launch post on X","url":"https://x.com/i/status/2118046765857546733","date":"2026-11-04"},{"kind":"repo","title":"ricursiveAI/Ripple-65B-Preview model card on Hugging Face","url":"https://huggingface.co/ricursiveAI/Ripple-65B-Preview"},{"kind":"repo","title":"ricursiveAI/Ripple-65B-Base-Preview on Hugging Face","url":"https://huggingface.co/ricursiveAI/Ripple-65B-Base-Preview"}],"firstBenchmark":null},"contextSource":"openrouter","organization":{"id":"ricursive","name":"Ricursive Intelligence","slug":"ricursive","country":null,"url":"https://www.ricursive.com","color":"#11110F","mark":"R"},"price":{"modelId":"ripple-65b-preview","provider":"ricursiveai","inputPerM":null,"outputPerM":null,"blendedPerM":0.15,"tokensPerS":null,"observedAt":"2026-11-22","sourceId":"openrouter"},"answer":{"word":"Pending benchmarks.","scope":"Announced Nov 4, 2026; no independent results yet.","lead":"As of Nov 22, 2026, Ripple-65B Preview is pending benchmarks on isitsota: announced Nov 4, 2026, with no independent results on the boards we track yet.","confidence":"low"},"status":"PENDING_BENCHMARKS","verdicts":[]}