YHK - yet another hacker news clone

Ripple-65B Preview: U.S. Open-Weight SOTA on Agentic Coding

130 points Philpax 2 hours ago 47 comments

wmf 1 hour ago | parent

Congratulations I guess but it's a pretty bad look to not even mention those other open weight models.

flatline 1 hour ago | parent

Right? These are not likely to be meaningful points of comparison to anyone but the authors.

wmf 1 hour ago | parent

Maybe Palantir will use it or something. There probably is a market for US models but I don't know how large it will be.

Onawa 1 hour ago | parent

US Gov won't let Chinese models be hosted on any internal infra, not even supposed to use them for local AI. So there is absolutely a market for US-made or non-Chinese models there.

tedivm 1 hour ago | parent

We've always been at war with eurasia.

Seriously though, it's just easier to pretend that China doesn't exist when they're stomping us this badly.

karmakaze 1 hour ago | parent

What's funnier is that after cherry picking models for comparison, they further cherry picked Inking for size comparison instead of better and smaller Beam 1.0

The bar chart also shows Nemotron 3 Ultra instead of smaller and better Nemotron 3.5 Super.

petcat 53 minutes ago | parent

Do those other models usually mention all the competition? I haven't noticed

stewartpark 43 minutes ago | parent

I think it's because the primary users of this model won't even look at "those other open weight models" as options. This is the right chart to present to those who would consider this model. (e.g. US Gov't agencies, chip design industry, etc etc)

schleck8 30 minutes ago | parent

Because Chinese models heavily rely on distillation and regularly behave strangely in public benchmarks (where they score much better than in private evaluations), even if many world class researchers are Chinese. It's difficult to quantify to what extend labs like Moonshot would be competitive if they could not distill from OpenAI, now that Anthropic mostly blocks it

https://www.theinformation.com/articles/openais-astra-model-...

RGS1811 1 hour ago | parent

The comparison chart is befuddling.

dofm 1 hour ago | parent

So we have a Recursive and a Ricursive, both making LLMs?

xyzrt 1 hour ago | parent

One making self-improving AI via AI-chip co-design, the other relying only on software for self-improving AI.

dofm 1 hour ago | parent

I hope they never have to call each other about anything. It could get confusing if neither knows who is the base case.

giwook 1 hour ago | parent

There seems to be some recursion going on here.

wmf 1 hour ago | parent

Time to start Rickursive, the lab that will never give you up and never let you down.

throwaw12 1 hour ago | parent

we need RichUrsive

6thbit 1 hour ago | parent

And it should acquire the existing two

neutronicus 58 minutes ago | parent

Roll them up, you might say.

syntaxing 1 hour ago | parent

When a post doesn’t mention Qwen for comparison, you know the performance isn’t going to be great.

ethan-boneh 1 hour ago | parent

I mean Qwen 3.8 27B hits 73% on TB2.1 and 42.2% on DeepSWE....

2001zhaozhao 1 hour ago | parent

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?

advael 1 hour ago | parent

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!

visarga 1 hour ago | parent

No, the term is well defined. It means being on the Pareto curve of performance vs size, such as: https://tinyurl.com/26sw35fw

O_H_E 1 hour ago | parent

well yes but the y axis heavily depends on the benchmark you choose

advael 51 minutes ago | parent

Irrelevant. A term can be well-defined and used correctly while also being diluted. In this case, I am pointing out that it's diluted by narrowing the category in which the claim is being made. I have no doubt that in this narrowed category, the term is being used correctly. This has no bearing on the fact that by narrowing the category in that way, this claim is made remarkably less impressive or interesting to many people hearing it, including myself

epolanski 32 minutes ago | parent

By...whom? An ISO standard? Highly cited scientific literature?

A tinyurl pic linked by randoms over the internet?

frinxor 49 minutes ago | parent

The US bit got me as well.

I guess it makes sense - there is a segmentation where users may want to run on domestic "frontier LLMs."

yorwba 32 minutes ago | parent

AFAIK the currently biggest US open-weight model is Inkling, which they include in the comparison, so at least their claim is not limited to a size category. Still they score 0.42 percentage points higher than the second-best model on their chosen benchmark, so it's hardly a meaningful lead.

gpm 28 minutes ago | parent

Playing games like this also makes it very easy to accidentally say things that aren't true, like in the headline of the article...

> Introducing Ripple-65B Preview, a Frontier Open-Weight Model

It's not "frontier", it's not "frontier" even in the category of open-weight models (though usually that phrasing would mean frontier over all, and also open weight)... it's only "frontier" in the small sub-category of US made open-weight models.

dev_l1x_be 26 minutes ago | parent

State of the AI-Apocalypse.

swiftcoder 20 minutes ago | parent

For sure, but unfortunately, there's a non-zero chance the state department at some point makes the "US" part of this mandatory for a bunch of companies.

jszymborski 1 hour ago | parent

Would love a 27B model. Wonder if we'll get a distillation.

6thbit 1 hour ago | parent

I'm sorry but what is "U.S." implying here?

Is this sponsored by the us gov? did it only train on US data? does it only work on the US?

conception 1 hour ago | parent

It's a US group that made it, in the same vein that OpenAI and Anthropic have US models.

petcat 1 hour ago | parent

It means that it's an American AI lab behind it.

mirekrusin 6 minutes ago | parent

...hiring Chinese people to do it.

signatoremo 34 minutes ago | parent

I think it’s the same as EU sovereign ones?

soundworlds 1 hour ago | parent

Not to be confused with https://recursive.com/

johnnyApplePRNG 1 hour ago | parent

Only benchmarking against other garbage american models?

LMFAO get out of my feed

stewartpark 58 minutes ago | parent

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!

regularfry 30 minutes ago | parent

What'll be interesting is how it performs relative to Qwen3.8 27b when quantised to the same size. Which, of course, is a comparison they don't want to do.

40four 22 minutes ago | parent

Well, obviously. These newer American models aren’t competing with Qwen yet. They are competing for American businesses and organizations that can’t/won’t use Chinese models.

They are playing catch-up, but I’m happy to see new players to the game, and it will be interesting to watch how they develop

59nadir 28 minutes ago | parent

Is it competitive overall or just competitive when compared to "U.S. Open-Weight models that are otherwise pretty decent at agentic coding"? The latter does not imply the former.

purplemantis 37 minutes ago | parent

These are pretty odd benchmarks to use. Terminal Bench 2.1 is old. SWE-bench was disavowed by OpenAI. DeepSWE is clearly benchmark-hacked by Gemini.

bluejay2387 23 minutes ago | parent

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.

minimaxa 17 minutes ago | parent

Agentic coding benchmarks look great.

I think we will find the market is going to move into omni multi-modal models: text, moe, agentic tasks, voice, images, and video all in one LLM.

thesnarkitecht 14 minutes ago | parent

Speaking as someone who has High-Memory Mac Studios, it is great that there is finally a model that will run at usable speeds with high quality. In my experience, the sweet spot for Mac Clusters is around 200GB of weights, which this is far under. Congratulations to the recursive team for building an open model that will actually run!