OpenYourMind · vision language · mixture of experts

Qwopus3.5-122B-A10B-Kimi-K2.6-destill-healed-abliterated

OpenYourMind/Qwopus3.5-122B-A10B-Kimi-K2.6-destill-healed-abliterated

Qwopus3.5-122B-A10B-Kimi-K2.6-destill-healed-abliterated at Q4_K_M is exactly 75,842,332,192 bytes (70.63 GiB / 75.84 GB) — an effective 4.950 bits per weight, not the nominal 4. Its KV cache at 32K is 0.75 GiB.

From the file· summed from 1 file(s)From the file· KV per layer
Parameters
123B
total, not active
Architecture
qwen35moe
48 layers
Context
262,144
native (config.json)
License
other

Shipped quantizations

● exact bytes, summed from published files
QuantSize●Exact bytes●Effective bpw●Tensors●Publisher
Q2_K42.66 GiB45,810,400,0322.990—mradermacher
Q3_K_S50.30 GiB54,007,208,7363.525—mradermacher
IQ2_M55.04 GiB59,097,948,6723.857—morikomorizz
Q3_K_M55.70 GiB59,807,722,2723.904899mradermacher
Q3_K_L60.22 GiB64,660,204,3204.221—mradermacher
IQ3_M62.59 GiB67,201,344,0004.386—morikomorizz
IQ4_XS62.92 GiB67,557,309,2164.410899mradermacher
Q4_K_S66.19 GiB71,074,282,2724.639—mradermacher
Q4_K_M70.63 GiB75,842,332,1924.950899OpenYourMind
Q4_K_M70.63 GiB75,842,333,4724.950—mradermacher
IQ4_NL76.06 GiB81,671,692,8005.331—morikomorizz
Q5_K_S80.03 GiB85,932,538,6565.609—mradermacher
Q5_K_M82.62 GiB88,708,557,6005.790899mradermacher

KV cache by context

computed per layer
ContextKV cache (f16)●Flat formulaOverstated byFull / windowed / recurrent
4,0960.09 GiB0.38 GiB4.00×12 / 0 / 36
8,1920.19 GiB0.75 GiB4.00×12 / 0 / 36
16,3840.38 GiB1.50 GiB4.00×12 / 0 / 36
32,7680.75 GiB3.00 GiB4.00×12 / 0 / 36
65,5361.50 GiB6.00 GiB4.00×12 / 0 / 36
131,0723.00 GiB12.00 GiB4.00×12 / 0 / 36

36 of 48 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-A3Bvs Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTPvs Qwen3.5-9Bvs gemma-4-26B-A4B-itvs gemma-4-12B-itvs Qwythos-9B-Claude-Mythos-5-1MAll alternatives →Will it fit on my card? →

Will it run on your card?

full quant x context sweep
By VRAM:8GB12GB16GB24GB32GB48GB
Radeon RX 6500 XT 4GBGeForce RTX 3050 6GBGeForce RTX 5050 8GBGeForce RTX 3080 10GBGeForce RTX 2080 Ti 11GBGeForce RTX 5070 12GBGeForce RTX 5060 Ti 16GBApple M3 Pro 18GBGeForce RTX 3080 Ti 20GBGeForce RTX 5090 D V2 24GBGeForce RTX 5090 D 32GBApple M5 Max 36GBApple M5 Pro 48GBApple M5 Max 64GBApple M3 Ultra 96GBApple M5 Max 128GBApple M2 Ultra 192GBApple M3 Ultra 256GBApple M3 Ultra 512GB

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 64.21 GiB. The real file is 70.63 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json
Layers
48
Attention heads
32
KV heads
2
Head dim
256
Hidden size
3072
Vocab
248,320
Sliding window
none
SWA period
—
MLA
no
Experts
256
Experts per token
8
use_sliding_window
—

Questions people ask

How much VRAM does Qwopus3.5-122B-A10B-Kimi-K2.6-destill-healed-abliterated need?
Q4_K_M is exactly 75,842,332,192 bytes (70.63 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is Qwopus3.5-122B-A10B-Kimi-K2.6-destill-healed-abliterated's KV cache?
0.75 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 Qwopus3.5-122B-A10B-Kimi-K2.6-destill-healed-abliterated a mixture-of-experts model?
Yes — 256 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 Qwopus3.5-122B-A10B-Kimi-K2.6-destill-healed-abliterated 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.