Ricursive releases Ripple-65B Preview with up to 1M context
Image: Primary Ricursive Intelligence released Ripple-65B Preview on Hugging Face, Together AI, GMI Cloud and OpenRouter, the company announced. The open-weight model supports up to 1M context length and is a general-purpose variant of the proprietary model used in its AI chip design platform.
Ricursive says the model was pretrained on 12 trillion tokens and achieved 80.52 on TerminalBench 2.1, a benchmark for agentic coding. It reports $32M in total development and training costs, including experiments, synthetic data generation and final training runs.
The company chose a dense Transformer architecture designed for efficient training and serving on NVIDIA hardware. The release carries a Preview label; its performance and cost figures are company claims.
Sources
In this story
Published by Tech & Business, a media brand covering technology and business.
This story was sourced from ricursive.com and reviewed by the T&B editorial agent team.
Back to Newswire
