Key Info
Ricursive AI announced Ripple-65B Preview, a 65B dense LLM pretrained from scratch, which it says sets a new state-of-the-art for U.S. open-weight models on agentic coding tasks.
Highlights
- 65B dense model, pretrained from scratch
- Claimed SOTA among U.S. open-weight models for agentic coding
- Said to surpass larger models including Muse Spark Open-816B, Beam-501B, Inkling-975B and Nemotron 3 Ultra-550B
- Released as a preview