Model reference · open weights

k2-merged-3.5T

LLMs NousResearch Text gen 2 builds Its own licence terms 330 dl/mo

k2-merged-3.5T is an open-weight language model from NousResearch. k2-merged-3.5T-bf16 (BF16) weighs 6936 GB; it needs more than the reference configurations below.

What it is

Released byNousResearch
Released2025-09-12
Parameters3468.1B
VRAM6936 GB for the weights

What it runs on

Memory and cards for k2-merged-3.5T-bf16 (BF16)

6936 GBweights, file size
119 MBcache per 1K tokens
4.2 GBruntime overhead, at least
131,072 tokenscontext max
CardRequests at onceContext maxMemory
8K each32K each
RTX 3060 12 GB … 8× B200 180 GB
16 smaller cards
———
Memory needed at each load
Requests at once8K tokens each32K tokens each
16942 GB6944 GB
56945 GB6960 GB
86948 GB6972 GB
166956 GB7003 GB
326972 GB7065 GB
647003 GB7189 GB

One card, with vLLM's small-card settings.

Builds

Sizes, precisions & builds

BuildParamsPrecisionWeightsSmallest setup
k2-merged-3.5T-bf16 (above) ↗ 3468.1BBF16 6936 GBnone of the reference configurations
k2-merged-3.5T-fp8 ↗ —FP8 3472 GBnone of the reference configurations

From the model card

What NousResearch says about k2-merged-3.5T

Experimental frankenmerge of Kimi K2-07, 09 and Base

Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.

© 2026 AxForge · EU-hosted AI infrastructure Pricing Docs Trust Privacy Terms