Can I run LFM3 1.7B locally?
LFM3 1.7B by LiquidAI has 1.7B parameters. At 4-bit with an 8K context it needs about 1.8 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track. See all models that fit in 8 GB of VRAM.
Memory needed by quant and context length
GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from LiquidAI/LFM3-1.7B-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.
| Quant | Quality | 4K ctx | 8K ctx | 32K ctx | 128K ctx |
|---|---|---|---|---|---|
| 2-bit (Q2_K) | rough | 1.3 GB | 1.3 GB | 1.6 GB | 2.8 GB |
| 3-bit (Q3_K_M) | usable | 1.5 GB | 1.5 GB | 1.8 GB | 2.9 GB |
| 4-bit (Q4_K_M) • | good, most popular | 1.7 GB | 1.8 GB | 2.0 GB | 3.2 GB |
| 5-bit (Q5_K_M) • | very good | 1.9 GB | 1.9 GB | 2.2 GB | 3.3 GB |
| 6-bit (Q6_K) • | near lossless | 2.0 GB | 2.1 GB | 2.4 GB | 3.5 GB |
| 8-bit (Q8_0) • | effectively lossless | 2.4 GB | 2.5 GB | 2.7 GB | 3.9 GB |
| 16-bit (F16/BF16) • | original weights | 7.2 GB | 7.2 GB | 7.5 GB | 8.6 GB |
Which GPUs and Macs run LFM3 1.7B?
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).
| Hardware | Memory | Fits? | Best quant | Needs | Max tok/s (est.) | Details |
|---|---|---|---|---|---|---|
| RTX 3060 12GB | 12 GB | Fits | F16 | 7.2 GB | 52 | details |
| RTX 4060 Ti 16GB | 16 GB | Fits | F16 | 7.2 GB | 42 | details |
| RTX 3090 | 24 GB | Fits | F16 | 7.2 GB | 135 | details |
| RTX 4090 | 24 GB | Fits | F16 | 7.2 GB | 145 | details |
| RTX 5090 | 32 GB | Fits | F16 | 7.2 GB | 258 | details |
| RTX 5070 Ti | 16 GB | Fits | F16 | 7.2 GB | 129 | details |
| RX 7900 XTX | 24 GB | Fits | F16 | 7.2 GB | 138 | details |
| Mac M4 Pro 48GB | 48 GB | Fits | F16 | 7.2 GB | 39 | details |
| Mac M4 Max 128GB | 128 GB | Fits | F16 | 7.2 GB | 79 | details |
| Mac M3 Ultra 512GB | 512 GB | Fits | F16 | 7.2 GB | 118 | details |
| DGX Spark | 128 GB | Fits | F16 | 7.2 GB | 39 | details |
| RTX 4070 | 12 GB | Fits | F16 | 7.2 GB | 73 | details |
| RTX 5060 Ti 16GB | 16 GB | Fits | F16 | 7.2 GB | 65 | details |
| RTX 4060 | 8 GB | Fits | Q8_0 | 2.5 GB | 147 | details |
| Arc B580 | 12 GB | Fits | F16 | 7.2 GB | 66 | details |
| RTX 3060 Ti | 8 GB | Fits | Q8_0 | 2.5 GB | 243 | details |
| RTX 3070 | 8 GB | Fits | Q8_0 | 2.5 GB | 243 | details |
| RTX 3080 10GB | 10 GB | Fits | F16 | 7.2 GB | 110 | details |
| RTX 3080 12GB | 12 GB | Fits | F16 | 7.2 GB | 132 | details |
| RTX 3090 Ti | 24 GB | Fits | F16 | 7.2 GB | 145 | details |
| RTX 4060 Ti 8GB | 8 GB | Fits | Q8_0 | 2.5 GB | 156 | details |
| RTX 4070 Super | 12 GB | Fits | F16 | 7.2 GB | 73 | details |
| RTX 4070 Ti | 12 GB | Fits | F16 | 7.2 GB | 73 | details |
| RTX 4070 Ti Super | 16 GB | Fits | F16 | 7.2 GB | 97 | details |
| RTX 4080 | 16 GB | Fits | F16 | 7.2 GB | 103 | details |
| RTX 4080 Super | 16 GB | Fits | F16 | 7.2 GB | 106 | details |
| RTX 5060 | 8 GB | Fits | Q8_0 | 2.5 GB | 243 | details |
| RTX 5070 | 12 GB | Fits | F16 | 7.2 GB | 97 | details |
| RTX 5080 | 16 GB | Fits | F16 | 7.2 GB | 138 | details |
| 2x RTX 3090 | 2×24 GB | Fits | F16 | 7.2 GB | 135 | details |
| 2x RTX 4090 | 2×24 GB | Fits | F16 | 7.2 GB | 145 | details |
| 2x RTX 5090 | 2×32 GB | Fits | F16 | 7.2 GB | 258 | details |
| RTX A6000 | 48 GB | Fits | F16 | 7.2 GB | 111 | details |
| RTX 6000 Ada | 48 GB | Fits | F16 | 7.2 GB | 138 | details |
| RTX PRO 6000 Blackwell | 96 GB | Fits | F16 | 7.2 GB | 258 | details |
| L40S | 48 GB | Fits | F16 | 7.2 GB | 125 | details |
| A100 40GB | 40 GB | Fits | F16 | 7.2 GB | 224 | details |
| A100 80GB | 80 GB | Fits | F16 | 7.2 GB | 294 | details |
| H100 PCIe | 80 GB | Fits | F16 | 7.2 GB | 288 | details |
| H100 SXM | 80 GB | Fits | F16 | 7.2 GB | 483 | details |
| H200 | 141 GB | Fits | F16 | 7.2 GB | 692 | details |
| RX 7600 | 8 GB | Fits | Q8_0 | 2.5 GB | 156 | details |
| RX 7700 XT | 12 GB | Fits | F16 | 7.2 GB | 62 | details |
| RX 7800 XT | 16 GB | Fits | F16 | 7.2 GB | 90 | details |
| RX 7900 GRE | 16 GB | Fits | F16 | 7.2 GB | 83 | details |
| RX 7900 XT | 20 GB | Fits | F16 | 7.2 GB | 115 | details |
| RX 9070 | 16 GB | Fits | F16 | 7.2 GB | 92 | details |
| RX 9070 XT | 16 GB | Fits | F16 | 7.2 GB | 92 | details |
| Ryzen AI Max+ 395 128GB | 128 GB | Fits | F16 | 7.2 GB | 37 | details |
| Arc A770 16GB | 16 GB | Fits | F16 | 7.2 GB | 81 | details |
| Mac M1 16GB | 16 GB | Fits | F16 | 7.2 GB | 9.8 | details |
| Mac M2 24GB | 24 GB | Fits | F16 | 7.2 GB | 14 | details |
| Mac M3 24GB | 24 GB | Fits | F16 | 7.2 GB | 14 | details |
| Mac M4 16GB | 16 GB | Fits | F16 | 7.2 GB | 17 | details |
| Mac M4 32GB | 32 GB | Fits | F16 | 7.2 GB | 17 | details |
| Mac M1 Pro 32GB | 32 GB | Fits | F16 | 7.2 GB | 29 | details |
| Mac M3 Pro 36GB | 36 GB | Fits | F16 | 7.2 GB | 22 | details |
| Mac M4 Pro 24GB | 24 GB | Fits | F16 | 7.2 GB | 39 | details |
| Mac M4 Pro 64GB | 64 GB | Fits | F16 | 7.2 GB | 39 | details |
| Mac M1 Max 64GB | 64 GB | Fits | F16 | 7.2 GB | 58 | details |
| Mac M2 Max 96GB | 96 GB | Fits | F16 | 7.2 GB | 58 | details |
| Mac M3 Max 128GB | 128 GB | Fits | F16 | 7.2 GB | 58 | details |
| Mac M4 Max 64GB | 64 GB | Fits | F16 | 7.2 GB | 79 | details |
| Mac M2 Ultra 192GB | 192 GB | Fits | F16 | 7.2 GB | 115 | details |
| Mac M3 Ultra 256GB | 256 GB | Fits | F16 | 7.2 GB | 118 | details |
Prefer an API?
LiquidAI: LFM3-1.7B is on OpenRouter at $0.08 per 1M input tokens and $0.20 per 1M output tokens. Price details and history →
Questions
How much VRAM does LFM3 1.7B need?
About 1.8 GB at 4-bit (Q4_K_M) with an 8K context: 1.0 GB of weights (real GGUF file size), 0.1 GB of KV cache and 0.6 GB of runtime overhead. At 8-bit it needs about 2.5 GB.
What is the smallest GPU that runs LFM3 1.7B?
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 LFM3 1.7B 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.
Similar models
Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.
LFM3 1.7B Base
1.7B · 1.8 GB at 4-bitd1 2B
1.9B · 1.8 GB at 4-bitLFM3 VL 2B
1.9B · 1.8 GB at 4-bitSmolLM2 1.7B Instruct
1.7B · 3.0 GB at 4-bitSmolLM2 1.7B
1.7B · 3.0 GB at 4-bitSmolLM 1.7B
1.7B · 3.0 GB at 4-bitEmbed this
Free badge and "Will it run on my GPU?" widget for model cards, READMEs and blogs. No cookies, no tracking. All options.
Copy a snippet
Markdown badge
[](https://runsonmy.com/model/lfm3-1-7b/) Widget (script)
<div class="runsonmy-widget" data-model="lfm3-1-7b"><a href="https://runsonmy.com/model/lfm3-1-7b/">Will LFM3 1.7B run on my GPU? Check on RunsOnMy</a></div>
<script async src="https://runsonmy.com/embed.js"></script> Widget (iframe)
<iframe src="https://runsonmy.com/embed/m/lfm3-1-7b/" width="100%" height="220" style="max-width:440px;border:0" loading="lazy" title="Will LFM3 1.7B run on my GPU? (RunsOnMy)"></iframe> Model facts
| Hugging Face repo | LiquidAI/LFM3-1.7B |
| Parameters | 1,728,801,536 (1.7B) |
| Layers / hidden size | 24 / 2048 |
| Attention | GQA + sliding window (hybrid: 13 linear/state layers); 16 heads, 4 KV heads, head dim 128; sliding window 256 |
| KV cache per token (FP16) | 22.0 KiB (sliding layers capped at the window) |
| Max context | 262,144 tokens |
| License | other |
| Released on HF | Oct 23, 2026 |
| Downloads (30d) / likes | 109,972 / 1,911 |
Sources: Hugging Face model API · config.json (LiquidAI/LFM3-1.7B) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.