Ricursive releases Ripple-65B Preview for agentic coding
Ricursive released Ripple-65B Preview, a 65B-parameter open-weight model pretrained on 12 trillion tokens with support for contexts up to 1 million tokens. The company says the model scored 80.52 on TerminalBench 2.1 and that its full research and development cost was $32 million. It is available through Hugging Face, Together AI, GMI Cloud and OpenRouter, with Fireworks listed as launching later that day.
Ricursive says Ripple-65B Preview combines leading reported performance among U.S. open-weight models on agentic coding with a relatively small model size and low reported development cost, making it directly available for local and hosted use.
Introducing Ripple-65B Preview,
a Frontier Open-Weight Model
Achieving U.S. Open-Weight SOTA on Agentic Coding
Ricursive Intelligence is a full-stack artificial intelligence lab building self-improving systems by co-optimizing hardware, software, and model.
To power our chip design platform, we trained a language model from scratch that sets a new U.S. open-weight SOTA on agentic coding. We optimized the model architecture for our training and serving platforms, driving substantial efficiency gains.
#Ripple-65B Preview#Ricursive Intelligence#TerminalBench 2.1#Reflection Beam 501B#Meta Muse Spark Open 816B#Thinking Machines Inkling 975B#NVIDIA Nemotron 3 Ultra 550B#Hugging Face#Together AI#GMI Cloud#OpenRouter#Fireworks