Ricursive trains a 65B open-weight model for $32M and beats rivals 10x its size
Ripple-65B Preview: U.S. Open-Weight SOTA on Agentic Coding
Ricursive Intelligence released Ripple-65B Preview, a dense Transformer trained from scratch on 12 trillion tokens with up to 1M context. It hits 80.52 on TerminalBench 2.1, edging out Reflection Beam 501B and far larger U.S. open-weight models. The full R&D run cost just $32M, roughly a tenth of the prior open-weight SOTA, and the model runs on local hardware like NVIDIA DGX Spark.
We reached a TerminalBench 2.1 score of 80.5 with only 1.4M rollouts, roughly 70x fewer rollouts than Reflection Beam, the prior SOTA.
- advael
Category narrowing to this degree really has diluted the term "SOTA" quite a lot
You mean it's state of the art for this subset of benchmarks you're targeting? No? Well surely it's state of the art for models of a certain size category? Huh, what about models that are available for local inference? O, state of the art for open models of a certain size made in the US specifically. Uh. Neat!
- stewartpark
It's great to see another competitive open weight model. Congratulations!
But as is, it's probably hard to become popular amongst hobbyists and local model enthusiasts like Qwen's 27B class models or MOE models because it's a 65B dense model and inference will be prohibitively slow on consumer grade hardware, and seems like the goal/mission of the company behind this model is to automate chip design.
So if I have to assume and connect the dots, this is for the companies who want to use AI in air-gapped environments for chip design so that they protect their IPs (not leaking to third party AI labs/hostile parties) and can afford lots of server grade inference hardware for a quick unlock with the big dense model.
Neat!
- wmf
Congratulations I guess but it's a pretty bad look to not even mention those other open weight models.
- 2001zhaozhao
This model scores 80.5 on Terminal Bench 2.1
For reference Qwen 3.8 Flash Next (which is bigger than this model but probably cheaper to run due to MoE) scores 84.3, and Qwen 3.8-27B scores 73.0.
I think this model could be competitive, depending on how much it was benchmaxxed compared to Qwen?
- bluejay2387
There are several situations where some of us are explicitly prohibited from using models from companies not under the jurisdiction of the United States. It doesn't mean these models are competitive when looked at from a global view but it is also not an irrelevant differentiation.