FRONTIER LAB / LLM · WORTH A LOOK

Ripple (Ricursive Intelligence)

65B open-weight coding model from Ricursive Intelligence (Apache 2.0, up to 1M context).

Visit ricursive.com ↗@RicursiveAI on X ↗

What Ripple (Ricursive Intelligence) does

Ripple-65B Preview is a dense 65-billion-parameter open-weight model from Ricursive Intelligence, released under Apache 2.0 with up to 1M context and pretrained on about 12T tokens. The company reports an 80.5 score on Terminal-Bench 2.1 and says it leads US open models on that test; those figures are Ricursive's own runs. Weights are on Hugging Face (ricursiveAI/Ripple-65B-Preview).

Good for

Self-hosted or API coding agents
Comparing open US coding models on your own tasks
Local or private inference where open weights matter

Try this in 10 minutes

Honest take

No red flags in our notes so far

Headline Terminal-Bench and other scores are Ricursive's own runs and have not been independently reproduced.
On some knowledge and coding rows larger rivals still lead in the company's own table.
No Sara or NoCodeLab mention in public bios.

Our Hype or Fact? checks

Partly factRipple-65B Preview: 'US open coding lead' on Terminal-Bench 5 Nov

Pricing

Open weights (Apache 2.0); hosting and inference are your cost. Ricursive also sells a proprietary chip-design platform.

Prices as listed by the maker when we last checked (2026-11-05); usually in US dollars. Check the website for current prices.

Latest update

5 November 2026: Released Ripple-65B Preview as open weights on Hugging Face; company claims 80.5 on Terminal-Bench 2.1.