RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

Can I run GLM 4.7 Flash locally?

GLM 4.7 Flash by zai-org has 31B parameters (4.6B active per token, 64 experts, 4 used per token). At 4-bit with an 8K context it needs about 18.2 GB of GPU memory. It fits on 48 of 65 GPUs and Macs we track.

18.2 GBQ4_K_M · 8K context
RTX 3090smallest consumer GPU that fits
M4 32GBsmallest Mac that fits
198Kmax context
MLAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from unsloth/GLM-4.7-Flash-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K) •rough11.5 GB11.7 GB12.9 GB17.9 GB
3-bit (Q3_K_M) •usable14.5 GB14.7 GB16.0 GB20.9 GB
4-bit (Q4_K_M) •good, most popular18.0 GB18.2 GB19.4 GB24.4 GB
5-bit (Q5_K_M) •very good20.8 GB21.0 GB22.3 GB27.2 GB
6-bit (Q6_K) •near lossless23.9 GB24.1 GB25.3 GB30.3 GB
8-bit (Q8_0) •effectively lossless30.6 GB30.8 GB32.0 GB37.0 GB
16-bit (F16/BF16) •original weights56.7 GB56.9 GB58.1 GB63.1 GB

Which GPUs and Macs run GLM 4.7 Flash?

Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).

HardwareMemoryFits?Best quantNeedsMax tok/s (est.)
RTX 3060 12GB12 GBNo———details
RTX 4060 Ti 16GB16 GBFitsQ2_K11.7 GB172details
RTX 309024 GBFitsQ4_K_M18.2 GB346details
RTX 409024 GBFitsQ4_K_M18.2 GB373details
RTX 509032 GBFitsQ6_K24.1 GB492details
RTX 5070 Ti16 GBFitsQ2_K11.7 GB535details
RX 7900 XTX24 GBFitsQ4_K_M18.2 GB355details
Mac M4 Pro 48GB48 GBFitsQ6_K24.1 GB75details
Mac M4 Max 128GB128 GBFitsF1656.9 GB62details
Mac M3 Ultra 512GB512 GBFitsF1656.9 GB93details
DGX Spark128 GBFitsF1656.9 GB31details
RTX 407012 GBNo———details
RTX 5060 Ti 16GB16 GBFitsQ2_K11.7 GB268details
RTX 40608 GBNo———details
Arc B58012 GBNo———details
RTX 3060 Ti8 GBNo———details
RTX 30708 GBNo———details
RTX 3080 10GB10 GBNo———details
RTX 3080 12GB12 GBNo———details
RTX 3090 Ti24 GBFitsQ4_K_M18.2 GB373details
RTX 4060 Ti 8GB8 GBNo———details
RTX 4070 Super12 GBNo———details
RTX 4070 Ti12 GBNo———details
RTX 4070 Ti Super16 GBFitsQ2_K11.7 GB401details
RTX 408016 GBFitsQ2_K11.7 GB428details
RTX 4080 Super16 GBFitsQ2_K11.7 GB440details
RTX 50608 GBNo———details
RTX 507012 GBNo———details
RTX 508016 GBFitsQ2_K11.7 GB573details
2x RTX 30902×24 GBFitsQ8_030.8 GB199details
2x RTX 40902×24 GBFitsQ8_030.8 GB214details
2x RTX 50902×32 GBFitsQ8_030.8 GB381details
RTX A600048 GBFitsQ8_030.8 GB163details
RTX 6000 Ada48 GBFitsQ8_030.8 GB204details
RTX PRO 6000 Blackwell96 GBFitsF1656.9 GB203details
L40S48 GBFitsQ8_030.8 GB184details
A100 40GB40 GBFitsQ8_030.8 GB331details
A100 80GB80 GBFitsF1656.9 GB231details
H100 PCIe80 GBFitsF1656.9 GB226details
H100 SXM80 GBFitsF1656.9 GB379details
H200141 GBFitsF1656.9 GB543details
RX 76008 GBNo———details
RX 7700 XT12 GBNo———details
RX 7800 XT16 GBFitsQ2_K11.7 GB373details
RX 7900 GRE16 GBFitsQ2_K11.7 GB344details
RX 7900 XT20 GBFitsQ3_K_M14.7 GB371details
RX 907016 GBFitsQ2_K11.7 GB382details
RX 9070 XT16 GBFitsQ2_K11.7 GB382details
Ryzen AI Max+ 395 128GB128 GBFitsF1656.9 GB29details
Arc A770 16GB16 GBFitsQ2_K11.7 GB334details
Mac M1 16GB16 GBNo———details
Mac M2 24GB24 GBFitsQ2_K11.7 GB60details
Mac M3 24GB24 GBFitsQ2_K11.7 GB60details
Mac M4 16GB16 GBNo———details
Mac M4 32GB32 GBFitsQ4_K_M18.2 GB44details
Mac M1 Pro 32GB32 GBFitsQ4_K_M18.2 GB74details
Mac M3 Pro 36GB36 GBFitsQ5_K_M21.0 GB47details
Mac M4 Pro 24GB24 GBFitsQ2_K11.7 GB163details
Mac M4 Pro 64GB64 GBFitsQ8_030.8 GB58details
Mac M1 Max 64GB64 GBFitsQ8_030.8 GB85details
Mac M2 Max 96GB96 GBFitsF1656.9 GB45details
Mac M3 Max 128GB128 GBFitsF1656.9 GB45details
Mac M4 Max 64GB64 GBFitsQ8_030.8 GB116details
Mac M2 Ultra 192GB192 GBFitsF1656.9 GB90details
Mac M3 Ultra 256GB256 GBFitsF1656.9 GB93details

Prefer an API?

Z.ai: GLM 4.7 Flash is on OpenRouter at $0.0605 per 1M input tokens and $0.40 per 1M output tokens. Price details and history →

Questions

How much VRAM does GLM 4.7 Flash need?

About 18.2 GB at 4-bit (Q4_K_M) with an 8K context: 17.1 GB of weights (real GGUF file size), 0.4 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 30.8 GB.

What is the smallest GPU that runs GLM 4.7 Flash?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 3090 (24 GB).

Can I run GLM 4.7 Flash on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M4 32GB or larger, using the default macOS GPU memory limit.

Model facts

Hugging Face repozai-org/GLM-4.7-Flash
Parameters31,221,488,576 (31B) · ~4.6B active (computed from config)
Layers / hidden size47 / 2048
AttentionMLA; 20 heads, 20 KV heads, head dim 102
KV cache per token (FP16)52.9 KiB
Max context202,752 tokens
Licensemit
Released on HFJan 19, 2026
Downloads (30d) / likes1,136,709 / 1,904

Sources: Hugging Face model API · config.json (zai-org/GLM-4.7-Flash) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.