RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

Can I run Maxwell V1 Flash locally?

Maxwell V1 Flash by maxwell-model has 37B parameters (2.3B active per token, 256 experts, 12 used per token). At 4-bit with an 8K context it needs about 22.1 GB of GPU memory. It fits on 48 of 65 GPUs and Macs we track.

22.1 GBQ4_K_M · 8K context
RTX 5090smallest consumer GPU that fits
M4 Pro 48GBsmallest Mac that fits
512Kmax context
GQAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K)rough13.8 GB14.0 GB15.0 GB19.3 GB
3-bit (Q3_K_M)usable17.7 GB17.9 GB18.9 GB23.2 GB
4-bit (Q4_K_M)good, most popular21.9 GB22.1 GB23.1 GB27.4 GB
5-bit (Q5_K_M)very good25.4 GB25.6 GB26.6 GB30.8 GB
6-bit (Q6_K)near lossless29.1 GB29.2 GB30.3 GB34.5 GB
8-bit (Q8_0)effectively lossless37.4 GB37.6 GB38.6 GB42.8 GB
16-bit (F16/BF16)original weights69.5 GB69.7 GB70.8 GB75.0 GB

Which GPUs and Macs run Maxwell V1 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 GBTightQ2_K14.0 GB338details
RTX 309024 GBFitsQ3_K_M17.9 GB843details
RTX 409024 GBFitsQ3_K_M17.9 GB908details
RTX 509032 GBFitsQ5_K_M25.6 GB1108details
RTX 5070 Ti16 GBTightQ2_K14.0 GB1052details
RX 7900 XTX24 GBFitsQ3_K_M17.9 GB865details
Mac M4 Pro 48GB48 GBFitsQ5_K_M25.6 GB169details
Mac M4 Max 128GB128 GBFitsF1669.7 GB120details
Mac M3 Ultra 512GB512 GBFitsF1669.7 GB180details
DGX Spark128 GBFitsF1669.7 GB60details
RTX 407012 GBNo———details
RTX 5060 Ti 16GB16 GBTightQ2_K14.0 GB526details
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 GBFitsQ3_K_M17.9 GB908details
RTX 4060 Ti 8GB8 GBNo———details
RTX 4070 Super12 GBNo———details
RTX 4070 Ti12 GBNo———details
RTX 4070 Ti Super16 GBTightQ2_K14.0 GB789details
RTX 408016 GBTightQ2_K14.0 GB842details
RTX 4080 Super16 GBTightQ2_K14.0 GB864details
RTX 50608 GBNo———details
RTX 507012 GBNo———details
RTX 508016 GBTightQ2_K14.0 GB1128details
2x RTX 30902×24 GBFitsQ8_037.6 GB388details
2x RTX 40902×24 GBFitsQ8_037.6 GB418details
2x RTX 50902×32 GBFitsQ8_037.6 GB743details
RTX A600048 GBFitsQ8_037.6 GB318details
RTX 6000 Ada48 GBFitsQ8_037.6 GB398details
RTX PRO 6000 Blackwell96 GBFitsF1669.7 GB395details
L40S48 GBFitsQ8_037.6 GB358details
A100 40GB40 GBFitsQ6_K29.2 GB835details
A100 80GB80 GBFitsQ8_037.6 GB845details
H100 PCIe80 GBFitsQ8_037.6 GB829details
H100 SXM80 GBFitsQ8_037.6 GB1389details
H200141 GBFitsF1669.7 GB1057details
RX 76008 GBNo———details
RX 7700 XT12 GBNo———details
RX 7800 XT16 GBTightQ2_K14.0 GB733details
RX 7900 GRE16 GBTightQ2_K14.0 GB677details
RX 7900 XT20 GBFitsQ2_K14.0 GB940details
RX 907016 GBTightQ2_K14.0 GB752details
RX 9070 XT16 GBTightQ2_K14.0 GB752details
Ryzen AI Max+ 395 128GB128 GBFitsF1669.7 GB56details
Arc A770 16GB16 GBTightQ2_K14.0 GB658details
Mac M1 16GB16 GBNo———details
Mac M2 24GB24 GBFitsQ2_K14.0 GB117details
Mac M3 24GB24 GBFitsQ2_K14.0 GB117details
Mac M4 16GB16 GBNo———details
Mac M4 32GB32 GBFitsQ3_K_M17.9 GB108details
Mac M1 Pro 32GB32 GBFitsQ3_K_M17.9 GB180details
Mac M3 Pro 36GB36 GBFitsQ3_K_M17.9 GB135details
Mac M4 Pro 24GB24 GBFitsQ2_K14.0 GB321details
Mac M4 Pro 64GB64 GBFitsQ8_037.6 GB113details
Mac M1 Max 64GB64 GBFitsQ8_037.6 GB166details
Mac M2 Max 96GB96 GBFitsQ8_037.6 GB166details
Mac M3 Max 128GB128 GBFitsF1669.7 GB88details
Mac M4 Max 64GB64 GBFitsQ8_037.6 GB226details
Mac M2 Ultra 192GB192 GBFitsF1669.7 GB176details
Mac M3 Ultra 256GB256 GBFitsF1669.7 GB180details

Questions

How much VRAM does Maxwell V1 Flash need?

About 22.1 GB at 4-bit (Q4_K_M) with an 8K context: 21.0 GB of weights (estimated), 0.4 GB of KV cache and 0.8 GB of runtime overhead. At 8-bit it needs about 37.6 GB.

What is the smallest GPU that runs Maxwell V1 Flash?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 5090 (32 GB).

Can I run Maxwell V1 Flash on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M4 Pro 48GB or larger, using the default macOS GPU memory limit.

Model facts

Hugging Face repomaxwell-model/Maxwell-V1-Flash
Parameters36,810,559,422 (37B) · ~2.3B active (computed from config)
Layers / hidden size45 / 2048
AttentionGQA; 32 heads, 4 KV heads, head dim 64
KV cache per token (FP16)45.0 KiB
Max context524,288 tokens
Licenseapache-2.0
Released on HFOct 13, 2026
Downloads (30d) / likes154 / 32

Sources: Hugging Face model API · config.json (maxwell-model/Maxwell-V1-Flash) · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.