LFM2 8B A1B — Hardware Requirements & GPU Compatibility
Chat
LFM2 8B A1B is Liquid AI's larger mixture-of-experts model, combining the company's novel hybrid architecture with approximately 8 billion total parameters. It uses a MoE design to keep active compute per token low while maintaining strong general performance across chat and reasoning tasks. For local users, it offers an intriguing alternative to conventional 8B transformers, with Liquid AI's architecture promising improved efficiency and throughput on consumer-grade hardware.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 125K)Best on 24 GB (Q4_K_M · ~350 tok/s · 125K)Macs from 8 GB (IQ3_M · ~64 tok/s)
Other
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run LFM2 8B A1B
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
ollama run hf.co/unsloth/LFM2-8B-A1B-GGUF:Q4_K_MQ4_K_M fits fully on the RTX 4090 · 18.5 GB free · 16K context
Download: LFM2-8B-A1B-Q4_K_M.gguf5.04 GBunsloth
Ways to runOther ways
each opens step-by-stepRun in cloud
Your own private copy · Q4_K_M · 24 GB GPU
- The easiest way to try this model
- Running in a few minutes, easy setup
- Chat right here or connect any app
$5+ bonus for new users
Runs on RTX 3060 12GB (6 GB headroom) · Q4_K_M
- Generation speed
- ~178 tok/s
- generation speed
- Cost per 1M output tokens
- $0.096
- per 1M output tokens
How Much VRAM Does LFM2 8B A1B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 4.0 GB | 4.3 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 125K context · ~409 tok/s | |
| 4.0 GB | 4.4 GB | 3-bit small — 3-bit small quantization | Fits: 125K context · ~404 tok/s | |
| 4.5 GB | 4.8 GB | 3-bit medium — 3-bit medium quantization | Fits: 125K context · ~386 tok/s | |
| 4.6 GB | 4.9 GB | 4-bit · legacy — 4-bit legacy quantization | Fits: 125K context · ~381 tok/s | |
| 5.4 GB | 5.7 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 125K context · ~350 tok/s | |
| 6.3 GB | 6.6 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 125K context · ~320 tok/s | |
| 7.3 GB | 7.6 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 125K context · ~295 tok/s | |
| 8.7 GB | 9 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 125K context · ~262 tok/s |
Selected Q4_K_M
4-bit medium quantization5.5GBat 16K context
- Weights
- 5.0 GB
- KV cache
- 0.2 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 18.5 GB free · ~350 tok/s
≥50 tok/s needs ≥83 GB/s · RTX 5090 ~458, RTX 3090 Ti ~350
Which Hardware Can Run LFM2 8B A1B?
Q4_K_M · 5.4 GBLFM2 8B A1B (Q4_K_M) requires 5.4 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 8+ GB is recommended. Using the full 128K context window can add up to 1.5 GB, bringing total usage to 6.9 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3070 Ti. 151 devices with unified memory can run LFM2 8B A1B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), NVIDIA Jetson Orin Nano 8GB (Super).
≤ 11 GB
≤ 11 GB tier →fits · 2.5 GB free · up to 125K ctx
16 GPUs · 1 device family
RTX 3080~298Simulate output on RTX 3080RTX 2080 Ti~261Simulate output on RTX 2080 TiRTX 3070 Ti~258Simulate output on RTX 3070 TiOrin Nano~61Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)20 GB
fits · 14.5 GB free · up to 125K ctx
1 GPU
RX 7900 XT~145Simulate output on RX 7900 XT32–40 GB
fits · 26.5 GB free · up to 125K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~458Simulate output on RTX 5090M5 Max36 GB~126Simulate output on Mac Studio M5 Max (36 GB)M532 GB~71Simulate output on MacBook Air 13" M5 (32 GB)Orin~112Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~72Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Alternatives to LFM2 8B A1B
LFM2 family · 9 models →Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Granite Switch 4.1 8BIBM | — | 8.4 GB | ~111 | same size |
| ALoDLM 8BAmazon | — | 7.6 GB | ~119 | same size |
| Llm Jp 4.1 8B Thinkingllm-jp | — | 7.6 GB | ~115 | same size |
| Ling 3.0 TinyInclusion AI | — | 5.5 GB | ~364 | same size |
| Apertus V1.5 8Bswiss-ai | — | 8.5 GB | ~111 | same size |
| BitCPM CANN 8BOpenBMB | — | 5.6 GB | ~127 | same size |
Specifications
- Publisher
- Liquid AI
- Family
- LFM2
- Parameters
- 8.3B
- Active parameters
- 1.6B per token
- Architecture
- Lfm2MoeForCausalLM
- Context Length
- 128,000 tokens
- Vocabulary Size
- 65,536
- Release Date
- 2025-10-07
- License
- Other
Frequently Asked Questions
How much VRAM does LFM2 8B A1B need?
LFM2 8B A1B requires 5.4 GB of VRAM at Q4_K_M, or 17.1 GB at BF16. Full 128K context adds up to 1.5 GB (6.9 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 8.3B × 4.8 bits ÷ 8 = 5 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 1.9 GB (at full 128K context)
VRAM usage by quantization
What's the best quantization for LFM2 8B A1B?
For LFM2 8B A1B, Q4_K_M (5.4 GB) offers the best balance of quality and VRAM usage. Q5_0 (5.6 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 2.7 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run LFM2 8B A1B on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run LFM2 8B A1B, the smallest at IQ3_M. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 5.4 GB of usable unified memory (RAM minus macOS overhead), from Mac Mini M2 Pro (16 GB).
Can I run LFM2 8B A1B locally?
Yes — LFM2 8B A1B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is LFM2 8B A1B?
At Q4_K_M, LFM2 8B A1B can reach ~200 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~350 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)
Mixture-of-Experts: just the active experts get read for each token, along with a fixed routing cost per layer.
Example: NVIDIA B200 → 1.0 GB active ÷ (8000 × 0.65) = 0.194 ms, plus 24 layers × 0.055 ms = 1.320 ms, so 1000 ÷ 1.514 ms = ~661 tok/s
Estimated speed at Q4_K_M (5.4 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of LFM2 8B A1B?
At Q4_K_M, the download is about 5.00 GB. The full-precision BF16 version is 16.68 GB. The smallest option (IQ2_XXS) is 2.29 GB.
Which GPUs can run LFM2 8B A1B?
52 consumer GPUs can run LFM2 8B A1B at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 7600. 40 GPUs have plenty of headroom for comfortable inference.
Which devices can run LFM2 8B A1B?
151 devices with unified memory can run LFM2 8B A1B at Q4_K_M (5.4 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.