Qwen3: Think deeper, act faster

Re:Qwen3: Think deeper, act faster by omneity on Monday April 28, 2025 @07:51PM (#43827359)

Excellent release by the Qwen team as always. Pretty much the best open-weights model line so far.

In my early tests however, several of the advertised languages are not really well supported and the model is outputting something that only barely resembles them.

Probably a dataset quality issue for low-resource languages that they cannot personally check for, despite the “119 languages and dialects” claim.

[ Parent | on HN ]

Re:Qwen3: Think deeper, act faster by jean- on Monday April 28, 2025 @09:55PM (#43828020)

Indeed, I tried several low-resource Romance languages they claim to support and performance is abysmal.

[ Parent | on HN ]

Re:Qwen3: Think deeper, act faster by vintermann on Tuesday April 29, 2025 @12:26AM (#43828681)

What size/quantification level? IME, small language performance is one of the things that really suffers from the various tricks that are used to reduce size.

[ Parent | on HN ]

Re:Qwen3: Think deeper, act faster by vitorgrs on Tuesday April 29, 2025 @01:55AM (#43829083)

Which languages?

[ Parent | on HN ]