How much VRAM does Qwen3.6-35B-A3B-REAM-192-heretic need?

IQ2_S is exactly 11,080,659,968 bytes (10.32 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

How large is Qwen3.6-35B-A3B-REAM-192-heretic's KV cache?

0.63 GiB at 32K context with an f16 cache, computed per layer from the model's architecture. Quantizing the cache to q8_0 roughly halves it.

Is Qwen3.6-35B-A3B-REAM-192-heretic a mixture-of-experts model?

Yes — 192 experts with 8 routed per token. Every expert must be resident in memory, but only the routed ones are read per token, so its memory requirement and its speed behave very differently.

keithnull · vision language · mixture of experts

Qwen3.6-35B-A3B-REAM-192-heretic

keithnull/Qwen3.6-35B-A3B-REAM-192-heretic

Qwen3.6-35B-A3B-REAM-192-heretic at IQ2_S is exactly 11,080,659,968 bytes (10.32 GiB / 11.08 GB) — an effective 3.277 bits per weight, not the nominal 2. Its KV cache at 32K is 0.63 GiB.

From the file· summed from 1 file(s)From the file· KV per layer

Parameters

27.0B

total, not active

Architecture

qwen35moe

40 layers

Context

262,144

native (config.json)

License

apache-2.0

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
IQ2_S	10.32 GiB	11,080,659,968	3.277	—	keithnull
Q3_K_L2 shards	24.91 GiB	26,742,828,800	7.910	—	keithnull
Q5_K_M2 shards	34.58 GiB	37,134,852,096	10.983	—	keithnull

KV cache by context

computed per layer

Context	KV cache (f16)●	Flat formula	Overstated by	Full / windowed / recurrent
4,096	0.08 GiB	0.31 GiB	4.00×	10 / 0 / 30
8,192	0.16 GiB	0.63 GiB	4.00×	10 / 0 / 30
16,384	0.31 GiB	1.25 GiB	4.00×	10 / 0 / 30
32,768	0.63 GiB	2.50 GiB	4.00×	10 / 0 / 30
65,536	1.25 GiB	5.00 GiB	4.00×	10 / 0 / 30
131,072	2.50 GiB	10.00 GiB	4.00×	10 / 0 / 30

30 of 40 layers use linear attention, which keeps a fixed-size recurrent state instead of a per-token cache. Those layers do not grow with context at all — treating them as ordinary attention, as a flat formula does, overstates this model's cache by roughly 4.0× at long context.

Compare with

same modality, comparable size

vs Qwen3.6-35B-A3B vs Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP vs Qwen3.5-9B vs gemma-4-26B-A4B-it vs gemma-4-12B-it vs Qwen3.5-4B All alternatives →Will it fit on my card? →

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts IQ2_S at roughly 14.17 GiB. The real file is 10.32 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json

Layers

Attention heads

KV heads

Head dim

256

Hidden size

2048

Vocab

248,320

Sliding window

none

SWA period

—

MLA

Experts

192

Experts per token

use_sliding_window

—

Questions people ask

How much VRAM does Qwen3.6-35B-A3B-REAM-192-heretic need?: IQ2_S is exactly 11,080,659,968 bytes (10.32 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is Qwen3.6-35B-A3B-REAM-192-heretic's KV cache?: 0.63 GiB at 32K context with an f16 cache, computed per layer. Quantizing the cache to q8_0 roughly halves it, which is often the difference between a context length fitting and not.
Is Qwen3.6-35B-A3B-REAM-192-heretic a mixture-of-experts model?: Yes — 192 experts, 8 routed per token. Every expert must be resident, but only the routed ones are read per token, which is why its memory requirement and its speed behave very differently.
Which quantization of Qwen3.6-35B-A3B-REAM-192-heretic should I use?: Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.