News for hackers. Stuff that matters.

A comment on DeepSeek-v3.2: Pushing the Frontier of Open Large Language Models [pdf]. Parent comment

Re:DeepSeek-v3.2: Pushing the Frontier of Open Large Languag... by halyconWays on Monday December 01, 2025 @02:42PM (#46112105)
As someone with a basement rig of 6x 3090s, not really. It's quite slow, as with that many params (685B) it's offloading basically all of it into system RAM. I limit myself to models with <144B params, then it's quite an enjoyable experience. GLM 4.5 Air has been great in particular
[ Parent | on HN ]
Re:DeepSeek-v3.2: Pushing the Frontier of Open Large Languag... by lostmsu on Tuesday December 02, 2025 @07:42AM (#46120653)
Did you find it better than GPT-OSS 120B? The public rankings are contradictory.
[ Parent | on HN ]
Re:DeepSeek-v3.2: Pushing the Frontier of Open Large Languag... by halyconWays on Sunday December 14, 2025 @07:52PM (#46268982)
I haven't used GPT-OSS 120B, or other GPT-OSS models, and I mostly go on personal recommendations rather than benchmarks directly.
[ Parent | on HN ]