Ripple-65B Preview

Model Nov 4, 2026

A dense 65B open-weight model with BF16, NVFP4 and base checkpoints; Ricursive reports strong agentic coding results and up to one million tokens of context.

Visit primary link 4 sources attached

Ricursive announcement video frame charts its reported Terminal-Bench 2.1 score against total model parameter counts. — Frame from Ricursive's original launch video. Benchmark scores and the smaller-model comparison are the author's claims; total parameters do not equate dense and mixture-of-experts inference compute. Source: AITrending browser capture / Ricursive Intelligence

coding-agents self-hosted

What happened lately

Ripple-65B Preview timeline

Ecosystem Nov 5, 2026

Ripple gets a 38.8 GB GGUF route—but stock llama.cpp cannot load it yet

Community reports 2 sources

Launches Nov 4, 2026

65B against the giants: Ricursive releases Ripple with an 80.52 coding claim

Community reports 4 sources

Ricursive announced Ripple-65B Preview on November 4, 2026. Its English model card describes a general-purpose dense model, Apache-2.0 weights and author-reported agentic coding benchmarks. Maximum-context hardware requirements and real task performance need evaluation for the intended configuration. This public release is distinct from the proprietary model used inside its chip-design platform.