zai-org · text · mixture of experts

GLM-4.7

zai-org/GLM-4.7

GLM-4.7 at Q4_K_M is exactly 216,455,572,576 bytes (201.59 GiB / 216.46 GB) — an effective 4.832 bits per weight, not the nominal 4. Its KV cache at 32K is 11.50 GiB.

From the file· summed from 5 file(s)From the file· KV per layer
Parameters
358B
total, not active
Architecture
glm4moe
92 layers
Context
202,752
native (config.json)
License
mit

Shipped quantizations

● exact bytes, summed from published files
QuantSize●Exact bytes●Effective bpw●Tensors●Publisher
IQ1_S2 shards71.50 GiB76,767,410,4641.714—bartowski
IQ1_M3 shards74.53 GiB80,024,529,3121.787—bartowski
UD-TQ1_078.69 GiB84,494,125,6641.886—unsloth
IQ2_XXS3 shards82.69 GiB88,785,566,0801.982—bartowski
UD-IQ1_S2 shards90.50 GiB97,175,341,8562.170—unsloth
IQ2_XS3 shards94.53 GiB101,503,891,8402.266—bartowski
IQ2_S3 shards95.29 GiB102,314,080,6722.284—bartowski
UD-IQ1_M3 shards100.27 GiB107,659,012,9922.404—unsloth
IQ2_M3 shards107.57 GiB115,504,265,6322.579—bartowski
UD-IQ2_XXS3 shards107.95 GiB115,912,821,6642.588—unsloth
UD-IQ2_M3 shards114.03 GiB122,440,207,2642.733—unsloth
Q2_K4 shards118.86 GiB127,623,776,8002.849—bartowski
Q2_K_L4 shards119.56 GiB128,381,536,8002.866—bartowski
Q2_K3 shards122.15 GiB131,159,649,1522.928—unsloth
Q2_K_L3 shards122.32 GiB131,341,511,5522.932—unsloth
IQ3_XXS4 shards132.93 GiB142,733,060,6403.187—bartowski
UD-IQ3_XXS3 shards135.15 GiB145,120,517,0243.240—unsloth
IQ3_XS4 shards138.48 GiB148,687,047,2003.320—bartowski
Q3_K_S4 shards144.38 GiB155,021,880,3203.461—unsloth
Q3_K_S4 shards146.52 GiB157,319,981,5683.512—bartowski
IQ3_M5 shards153.45 GiB164,764,174,9763.678—bartowski
Q3_K_M5 shards153.63 GiB164,954,966,6563.683—bartowski
Q3_K_L5 shards158.87 GiB170,584,017,5683.808—bartowski
Q3_K_M4 shards159.51 GiB171,268,787,2323.824—unsloth
IQ4_XS4 shards178.41 GiB191,562,951,7124.277—unsloth
IQ4_XS5 shards179.45 GiB192,678,005,3764.302—bartowski
IQ4_NL5 shards188.55 GiB202,456,345,7284.520—unsloth
Q4_05 shards189.09 GiB203,036,339,3284.533—unsloth
IQ4_NL6 shards189.62 GiB203,602,938,6244.545—bartowski
Q4_K_S5 shards189.69 GiB203,679,902,8484.547—unsloth
Q4_06 shards192.31 GiB206,494,059,2644.610—bartowski
Q4_K_S6 shards196.30 GiB210,778,802,9444.706—bartowski
Q4_K_M5 shards201.59 GiB216,455,572,5764.832—unsloth
Q4_K_M6 shards203.52 GiB218,524,687,1364.879—bartowski
Q4_K_L6 shards204.05 GiB219,100,584,7044.891—bartowski
Q4_15 shards209.18 GiB224,605,547,6485.014—unsloth
Q4_16 shards209.72 GiB225,189,350,1125.027—bartowski
Q5_K_S5 shards230.04 GiB247,003,704,4485.514—unsloth
Q5_K_S7 shards230.45 GiB247,443,676,0325.524—bartowski
Q5_K_M6 shards236.79 GiB254,246,825,2485.676—unsloth

KV cache by context

computed per layer
ContextKV cache (f16)●Flat formulaOverstated byFull / windowed / recurrent
4,0961.44 GiB1.44 GiB—92 / 0 / 0
8,1922.88 GiB2.88 GiB—92 / 0 / 0
16,3845.75 GiB5.75 GiB—92 / 0 / 0
32,76811.50 GiB11.50 GiB—92 / 0 / 0
65,53623.00 GiB23.00 GiB—92 / 0 / 0
131,07246.00 GiB46.00 GiB—92 / 0 / 0

Compare with

same modality, comparable size
vs Qwen3-Coder-30B-A3B-Instructvs Qwen3.6-27Bvs Qwen3.8-27Bvs DeepSeek-V4-Flashvs Qwen3-30B-A3B-Thinking-2507vs Laguna-XS-2.1All 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 187.72 GiB. The real file is 201.59 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json
Layers
92
Attention heads
96
KV heads
8
Head dim
128
Hidden size
5120
Vocab
151,552
Sliding window
none
SWA period
—
MLA
no
Experts
160
Experts per token
8
use_sliding_window
—

Questions people ask

How much VRAM does GLM-4.7 need?
Q4_K_M is exactly 216,455,572,576 bytes (201.59 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is GLM-4.7's KV cache?
11.50 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 GLM-4.7 a mixture-of-experts model?
Yes — 160 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 GLM-4.7 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.