anana_

Apparently agentic performance in Gemma was improved recently: https://x.com/googlegemma/status/2077449152062247219

Too little too late imo

They keep mentioning a 31B dense model, but there are no benchmarks or weights for it anywhere?

It looks like the purpose of this model is to i. generate environmental sim data for doing RL on other models or ii. act as a foundation model (they trained it to select actions as well as predicting the next state in the same loop?)

Either way, neither are intended for end consumers.

I believe the benchmark listed is about simulating the environment for the various tasks, rather than doing them. It seems that the point of this model is to generate sim data to improve other models with

Unfortunately on Strix Halo or any similar unified memory set up, dense models are gonna be dirt slow due to the tiny memory bandwidth... But I agree, 27B is superior.

Perhaps try a different model? Just from anecdotal experience, I find that the Gemma models smaller than 31B do not tool call as often as they should.

Some of the benchmarks appear to back this up [0]

Of course, a lot depends how you are using it (inference parameters, harness, prompting, etc.), but the model is quite important too.

[0] https://artificialanalysis.ai/models/open-source/small?model...

I have one too and it never occurred to me to use it for anything other than games. Would be interested in seeing how you did it!

They do now - https://support.mozilla.org/en-US/kb/use-sidebar-access-tool...

Hypothesizing here, but maybe the idea is sort of a form of technological/economic warfare? Releasing performance equivalent yet more cost efficient open weight models should in theory drive the cost of inference down everywhere.

This I assume will make it more difficult for US AI labs to turn a profit, which might make investors question their sky high valuations.

Any sort of melt down in the AI sector would almost certainly spread to the wider US market.

In contrast, in China, most of the funding for AI is coming directly from the government, so it's unlikely the same capital flight scenario would happen.

I'm not saying it's the latest Qwen iteration - that would be Qwen3.6.

I'm saying it's the latest iteration of the finetuned model mentioned in the parent comment.

I'm also not suggesting that it's "the latest and greatest" anything. In fact, I think it's rather clear that I'm suggesting the opposite? As in - how can a small fine tune produce better results than a frontier lab's work?

It's rather surprising that a solo dev can squeeze more performance out of a model with rather humble resources vs a frontier lab. I'm skeptical of claims that such a fine-tuned model is "better" -- maybe on certain benchmarks, but overall?

FYI the latest iteration of that finetune is here: https://huggingface.co/Jackrong/Qwopus3.5-27B-v3

Upon rereading, I'd agree. Fits with the tone of the rest of the write up.

Sometimes you need the absolute cutting-edge reasoning of Claude 3.5 Sonnet or GPT-4o

Dead giveaway

https://huggingface.co/Qwen/Qwen3.5-27B

I wasn't aware of that, which page mentions that?

I've had even better results using the dense 27B model -- less looping and churning on problems

They own GEICO...