News for hackers. Stuff that matters.

A comment on I'm making LLMs work on small and old devices. Parent comment

Re:I'm making LLMs work on small and old devices by yjftsjthsd-h on Thursday November 12, 2026 @02:56PM (#50473269)
I'm told that qwen 3.8 27B is good enough for practical programming work, so if that's doable then I could easily see the utility. Or... the biggest qwen I see here is Qwen3.5-2B; is that actually capable of programming or agentic work with code?
[ Parent | on HN ]
Re:I'm making LLMs work on small and old devices by Miraste on Thursday November 12, 2026 @03:34PM (#50473828)
[delayed]
[ Parent | on HN ]
Re:I'm making LLMs work on small and old devices by NooneAtAll5 on Thursday November 12, 2026 @07:45PM (#50476629)
> In the last month we've gotten 4B models that can do small tasks.

where can I read up on that?

[ Parent | on HN ]
Re:I'm making LLMs work on small and old devices by mrashad74 on Friday November 13, 2026 @01:01AM (#50478956)
Check LFM as a starting pointer : https://www.liquid.ai/models
[ Parent | on HN ]
Re:I'm making LLMs work on small and old devices by Miraste on Friday November 13, 2026 @10:17AM (#50484773)
The best one I know of at the moment is XHToken's Spark-x2.6 4B.
[ Parent | on HN ]