natrys

Can you perhaps keep the challenge up with some null flag (at least for a bit)? It would be nice to check if my locally validated solution works ;)

I think we don't have a good draft model for better speculative decoding yet (e.g. DFlash 2). Once we do, it will be faster.

Why not? It's not really competing in the same size class.

Besides, as they explicitly wrote here, the main goal for this release is not performance, rather to serve as a reference for inference runtimes about what to implement. So that later Qwen 4 can be released with zero day support.

For me, flash 0731 was much better in omp/opencode than in Pi.

Anyway, it might be so that they are rolling out deployment. There haven't been an official announcement post yet (this submission is a link to openrouter). Some people have been saying they are getting results worse than GLM-5.1, that's obviously broken.

There is an old rule of thumb that says the quality of an MoE is equivalent to a dense model with the geometric mean of its total and active parameters. So, for example, the Qwen3.6 would be equivalent to a dense model with approximately sqrt(35×3) ≈ 10.25B parameters.

Both MoEs and dense models are always getting better, so I don't think this comparison is meaningful across generations. But still for a first approximation, this tends to hold (you wouldn't expect a lot from a 10B model in coding yet).

So your cost would be $40.68 with another provider that has one less zero in the cache hit price.

At least the SSM layers should have relative PE built in via recurrence, in a hybrid attention model like this.

They can and almost certainly are doing similarly impressive engineering works internally.

They just aren't in any hurry to forward those cost savings to you.