Ricursive open-sources Ripple-65B Preview model
What happened
On 7 November 2026, AI chip design company Ricursive Intelligence trained and released the open-source language model Ripple-65B Preview under the Apache 2.0 licence. According to The Decoder, the model was pre-trained on 12 trillion tokens and uses a dense 65-billion-parameter Transformer architecture rather than MoE, optimised for Nvidia hardware and focused on agentic coding. The release is still at the launch stage, with no reports yet of evaluations or user feedback.
Written by AI from the reports · updated 42 min ago
Timeline
Follow the reports to see the event from each side.
- The DecoderRipple-65B Preview delivers strong agentic coding results as a fully Apache 2.0 open US model
Ricursive Intelligence is releasing Ripple-65B Preview, a 65-billion-parameter open-weight language model that the company says achieves the top score among open-source U.S. models on the TerminalBench 2.1 coding benchmark at a total R&D cost of $32 million. On other benchmarks, however, Ripple falls behind the likewise newly released western open model Inkling 2 Small.
Heat of this event
Not enough continuous observations yet to draw a trend.