Ripple-65B Preview posts strong coding benchmark results

Ricursive Intelligence has released Ripple-65B Preview, an open-weight language model with 65 billion parameters. The AI chip design company says it was pretrained on 12 trillion tokens and reached 80.52 on TerminalBench 2.1, which it describes as the highest score from an open U.S. model. The stated total cost for research, development and training was $32 million.

Ripple uses a harness based partly on Terminus 2 and partly on OpenCode. On several other evaluations, it scored below the newly released Inkling 2 Small: 79.1% versus 86.3% on SWE-Bench Pro v2.0 Hard, 44.3% versus 72.1% on DeepSWE v1.1, and 71.67% versus 85.4/85.0% on MCP Atlas. The model uses a dense Transformer design rather than a mixture-of-experts architecture and is optimized for Nvidia hardware. Ricursive says its structure could make it easier to run on users’ own hardware. Ripple-65B Preview is available through Hugging Face under the Apache 2.0 license.