HN Explorer

Ask HN: What's the best LLM model that on a 24 GB VRAM GPU?

3 points by max93 ↗ HN
What’s the best model right now that outperforms Qwopus3.6-27B-v2-MTP-GGUF 8-bit on a 24 GB VRAM GPU? Looking for real reviews. I found 4 bit not usable in production.

3 comments

[ 4.8 ms ] story [ 15.6 ms ] thread
roscas
I would love to see Laguna.xs-2 on that, because it's very good on cpu only.
jr_isidore
Good question. I was told in 2024 to get an RTX 3090 (24 GB VRAM), so I did, and nothing on HuggingFace was usable.
sds357
qwen3.6 27b and 35b work for me. I run them on proxmox with a an old P40 gpu, and ollama