Hermes405B params128k context
Hermes 3 405B locally
Hermes 3 405B is an open-weight LLM from the Hermes family with 405B parameters. Main use: chat, RAG and general assistance. Detected minimum hardware: Instinct MI325X (256 GB).
Technical facts
Parameters405B
Max context128k
Q4_K_M254.6 GB
Q5_K_M311.2 GB
Q8452.6 GB
FP16905.2 GB
FamilyHermes
Last sync2026-05-12
Available quantizations
GGUF weightsQ4_K_M
254.6GB
Acceptable. Good compromise when VRAM is limited.
Q5_K_M
311.2GB
Good quality. Sweet spot for size and precision.
Q8
452.6GB
Near-FP16 quality. Comfortable for production.
FP16
905.2GB
Reference precision. Maximum quality, doubled VRAM.
Compatible GPUs
3 single-GPUGPUs that can run Hermes 3 405B on a single card, ranked by VRAM headroom.
Similar models
VRAM estimates: parameters x bits/8 plus margin. Real performance varies by engine, context length and batch size.
sync: 2026-05-12