Ricursive Intelligence, an AI chip design firm, has trained an open language model called Ripple-65B Preview. The company says the model was pretrained on 12 trillion tokens.
According to Ricursive, Ripple-65B Preview scored 80.52 on the TerminalBench 2.1 coding benchmark. The company describes this as the best result among open US models. The model uses Terminus 2 and OpenCode as its harness, according to the report.
The report says other benchmark results show Ripple does not match Inkling 2 Small, a model also newly released. However, the report states Ripple is likely easier to deploy on users' own hardware.
Ricursive says total development and training costs came to $32 million, covering all research and development expenses.
Ripple-65B Preview uses a dense Transformer architecture with 65 billion parameters, rather than the more common mixture-of-experts approach. The model is optimized for Nvidia hardware. It is available on Hugging Face under the Apache 2.0 license.
Source: The Decoder