RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

Can I run LFM2.5 1.2B Instruct locally?

LFM2.5 1.2B Instruct by LiquidAI has 1.2B parameters. At 4-bit with an 8K context it needs about 1.3 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track.

1.3 GBQ4_K_M · 8K context
RTX 4060smallest consumer GPU that fits
M1 16GBsmallest Mac that fits
125Kmax context
GQAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from LiquidAI/LFM2.5-1.2B-Instruct-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx
2-bit (Q2_K)rough1.0 GB1.0 GB1.3 GB
3-bit (Q3_K_M)usable1.1 GB1.2 GB1.4 GB
4-bit (Q4_K_M) •good, most popular1.3 GB1.3 GB1.6 GB
5-bit (Q5_K_M) •very good1.4 GB1.4 GB1.7 GB
6-bit (Q6_K) •near lossless1.5 GB1.5 GB1.8 GB
8-bit (Q8_0) •effectively lossless1.7 GB1.8 GB2.1 GB
16-bit (F16/BF16) •original weights4.9 GB5.0 GB5.3 GB

Which GPUs and Macs run LFM2.5 1.2B Instruct?

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 GBFitsF165.0 GB77details
RTX 4060 Ti 16GB16 GBFitsF165.0 GB61details
RTX 309024 GBFitsF165.0 GB200details
RTX 409024 GBFitsF165.0 GB215details
RTX 509032 GBFitsF165.0 GB382details
RTX 5070 Ti16 GBFitsF165.0 GB191details
RX 7900 XTX24 GBFitsF165.0 GB205details
Mac M4 Pro 48GB48 GBFitsF165.0 GB58details
Mac M4 Max 128GB128 GBFitsF165.0 GB117details
Mac M3 Ultra 512GB512 GBFitsF165.0 GB175details
DGX Spark128 GBFitsF165.0 GB58details
RTX 407012 GBFitsF165.0 GB108details
RTX 5060 Ti 16GB16 GBFitsF165.0 GB96details
RTX 40608 GBFitsF165.0 GB58details
Arc B58012 GBFitsF165.0 GB97details
RTX 3060 Ti8 GBFitsF165.0 GB96details
RTX 30708 GBFitsF165.0 GB96details
RTX 3080 10GB10 GBFitsF165.0 GB162details
RTX 3080 12GB12 GBFitsF165.0 GB195details
RTX 3090 Ti24 GBFitsF165.0 GB215details
RTX 4060 Ti 8GB8 GBFitsF165.0 GB61details
RTX 4070 Super12 GBFitsF165.0 GB108details
RTX 4070 Ti12 GBFitsF165.0 GB108details
RTX 4070 Ti Super16 GBFitsF165.0 GB143details
RTX 408016 GBFitsF165.0 GB153details
RTX 4080 Super16 GBFitsF165.0 GB157details
RTX 50608 GBFitsF165.0 GB96details
RTX 507012 GBFitsF165.0 GB143details
RTX 508016 GBFitsF165.0 GB205details
2x RTX 30902×24 GBFitsF165.0 GB200details
2x RTX 40902×24 GBFitsF165.0 GB215details
2x RTX 50902×32 GBFitsF165.0 GB382details
RTX A600048 GBFitsF165.0 GB164details
RTX 6000 Ada48 GBFitsF165.0 GB205details
RTX PRO 6000 Blackwell96 GBFitsF165.0 GB382details
L40S48 GBFitsF165.0 GB184details
A100 40GB40 GBFitsF165.0 GB332details
A100 80GB80 GBFitsF165.0 GB435details
H100 PCIe80 GBFitsF165.0 GB427details
H100 SXM80 GBFitsF165.0 GB715details
H200141 GBFitsF165.0 GB1024details
RX 76008 GBFitsF165.0 GB61details
RX 7700 XT12 GBFitsF165.0 GB92details
RX 7800 XT16 GBFitsF165.0 GB133details
RX 7900 GRE16 GBFitsF165.0 GB123details
RX 7900 XT20 GBFitsF165.0 GB171details
RX 907016 GBFitsF165.0 GB137details
RX 9070 XT16 GBFitsF165.0 GB137details
Ryzen AI Max+ 395 128GB128 GBFitsF165.0 GB55details
Arc A770 16GB16 GBFitsF165.0 GB119details
Mac M1 16GB16 GBFitsF165.0 GB15details
Mac M2 24GB24 GBFitsF165.0 GB21details
Mac M3 24GB24 GBFitsF165.0 GB21details
Mac M4 16GB16 GBFitsF165.0 GB26details
Mac M4 32GB32 GBFitsF165.0 GB26details
Mac M1 Pro 32GB32 GBFitsF165.0 GB43details
Mac M3 Pro 36GB36 GBFitsF165.0 GB32details
Mac M4 Pro 24GB24 GBFitsF165.0 GB58details
Mac M4 Pro 64GB64 GBFitsF165.0 GB58details
Mac M1 Max 64GB64 GBFitsF165.0 GB85details
Mac M2 Max 96GB96 GBFitsF165.0 GB85details
Mac M3 Max 128GB128 GBFitsF165.0 GB85details
Mac M4 Max 64GB64 GBFitsF165.0 GB117details
Mac M2 Ultra 192GB192 GBFitsF165.0 GB171details
Mac M3 Ultra 256GB256 GBFitsF165.0 GB175details

Questions

How much VRAM does LFM2.5 1.2B Instruct need?

About 1.3 GB at 4-bit (Q4_K_M) with an 8K context: 0.7 GB of weights (real GGUF file size), 0.1 GB of KV cache and 0.5 GB of runtime overhead. At 8-bit it needs about 1.8 GB.

What is the smallest GPU that runs LFM2.5 1.2B Instruct?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 4060 (8 GB).

Can I run LFM2.5 1.2B Instruct on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M1 16GB or larger, using the default macOS GPU memory limit.

Model facts

Hugging Face repoLiquidAI/LFM2.5-1.2B-Instruct
Parameters1,170,340,608 (1.2B)
Layers / hidden size16 / 2048
AttentionGQA (hybrid: 10 linear/state layers); 32 heads, 8 KV heads, head dim 64
KV cache per token (FP16)12.0 KiB
Max context128,000 tokens
Licenseother
Released on HFJan 6, 2026
Downloads (30d) / likes105,811 / 687

Sources: Hugging Face model API · config.json (LiquidAI/LFM2.5-1.2B-Instruct) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.