LFM2 24B A2B — Hardware Requirements & GPU Compatibility
ChatFunctions
LFM2 24B A2B is a 23.8B-parameter open language model from Liquid AI in the LFM2 family. It supports a context window of up to 128,000 tokens. At Q4_K_M it needs about 14.77 GB of VRAM — see which GPUs and Macs can run it below.
Runs great on a 24 GB GPU
Runs from 12 GB (Q2_K · 64K)Best on 24 GB (Q4_K_M · ~227 tok/s · 125K)Macs from 16 GB (Q3_K_M · ~61 tok/s)
Other
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run LFM2 24B A2B
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 lfm2:24bQ4_K_M fits fully on the RTX 4090 · 9.1 GB free · 16K context
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3090 (24 GB) (9 GB headroom) · Q4_K_M
- Generation speed
- ~219 tok/s
- generation speed
- Cost per 1M output tokens
- $0.23
- per 1M output tokens
How Much VRAM Does LFM2 24B A2B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 10.6 GB | 11.1 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 125K context · ~265 tok/s | |
| 12.1 GB | 12.6 GB | 3-bit medium — 3-bit medium quantization | Fits: 125K context · ~250 tok/s | |
| 14.8 GB | 15.3 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 125K context · ~227 tok/s | |
| 17.5 GB | 18 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 125K context · ~208 tok/s | |
| 20.1 GB | 20.6 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 125K context · ~192 tok/s | |
| 24.3 GB | 24.8 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Does not fit: needs 24.5 GB | |
| 48.2 GB | 48.7 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 48.3 GB |
Selected Q4_K_M
4-bit medium quantization14.9GBat 16K context
- Weights
- 14.3 GB
- KV cache
- 0.3 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 9.1 GB free · ~227 tok/s
≥50 tok/s needs ≥124 GB/s · RTX 5090 ~291, RTX 3090 Ti ~227
Which Hardware Can Run LFM2 24B A2B?
Q4_K_M · 14.8 GBLFM2 24B A2B (Q4_K_M) requires 14.8 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 20+ GB is recommended. Using the full 128K context window can add up to 2.4 GB, bringing total usage to 17.2 GB. 26 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 5080. 122 devices with unified memory can run LFM2 24B A2B, including NVIDIA DGX Spark, NVIDIA Jetson AGX Thor Developer Kit, Mac Mini M5 Pro (24 GB).
16–18 GB
16–18 GB tier →fits · 1.1 GB free · up to 64K ctx
18 GPUs
RTX 5080~222Simulate output on RTX 5080RTX 5070 Ti~214Simulate output on RTX 5070 TiRTX 4080 SUPER~192Simulate output on RTX 4080 SUPER20 GB
fits · 5.1 GB free · up to 125K ctx
1 GPU
RX 7900 XT~91Simulate output on RX 7900 XT32–40 GB
fits · 17.1 GB free · up to 125K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~291Simulate output on RTX 5090M5 Max36 GB~80Simulate output on Mac Studio M5 Max (36 GB)M532 GB~47Simulate output on MacBook Air 13" M5 (32 GB)Orin~77Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~49Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Alternatives to LFM2 24B A2B
LFM2 family · 9 models →Similar models
4| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Magistral Small 2509Mistral AI | 85.6 | 18.1 GB | ~43 | same size |
| Ariarhymes-ai | — | 20.2 GB | ~168 | same size |
| EuroLLM 22B Instruct 2512utter-project | — | 17.5 GB | ~46 | same size |
| LFM2.5 24B A2BLiquid AI | — | 15 GB | ~221 | same size |
Specifications
- Publisher
- Liquid AI
- Family
- LFM2
- Parameters
- 23.8B
- Active parameters
- 2.3B per token
- Architecture
- Lfm2MoeForCausalLM
- Context Length
- 128,000 tokens
- Vocabulary Size
- 65,536
- Release Date
- 2026-02-24
- License
- Other
Frequently Asked Questions
How much VRAM does LFM2 24B A2B need?
LFM2 24B A2B requires 14.8 GB of VRAM at Q4_K_M, or 48.2 GB at BF16. Full 128K context adds up to 2.4 GB (17.2 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 23.8B × 4.8 bits ÷ 8 = 14.3 GB
KV Cache + Overhead ≈ 0.5 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 ≈ 2.9 GB (at full 128K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 4090 run LFM2 24B A2B?
Yes, at Q6_K (20.1 GB) or lower. Higher quantizations like Q8_0 (24.3 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
What's the best quantization for LFM2 24B A2B?
For LFM2 24B A2B, Q4_K_M (14.8 GB) offers the best balance of quality and VRAM usage. Q5_K_M (17.5 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 10.6 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run LFM2 24B A2B on a Mac?
Yes — Mac Mini M2 Pro (16 GB) and 133 other Macs can run LFM2 24B A2B, the smallest at Q3_K_M. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 14.8 GB of usable unified memory (RAM minus macOS overhead), from Mac Mini M5 Pro (24 GB).
Can I run LFM2 24B A2B locally?
Yes — LFM2 24B A2B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is LFM2 24B A2B?
At Q4_K_M, LFM2 24B A2B can reach ~121 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~227 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 → 1.4 GB active ÷ (8000 × 0.65) = 0.277 ms, plus 40 layers × 0.055 ms = 2.200 ms, so 1000 ÷ 2.477 ms = ~404 tok/s
Estimated speed at Q4_K_M (14.8 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of LFM2 24B A2B?
At Q4_K_M, the download is about 14.31 GB. The full-precision BF16 version is 47.69 GB. The smallest option (Q2_K) is 10.13 GB.
Which GPUs can run LFM2 24B A2B?
26 consumer GPUs can run LFM2 24B A2B at Q4_K_M. Top options include AMD Radeon RX 7900 XTX, NVIDIA GeForce RTX 3090, NVIDIA GeForce RTX 3090 Ti, AMD Radeon RX 6800. 7 GPUs have plenty of headroom for comfortable inference.
Which devices can run LFM2 24B A2B?
122 devices with unified memory can run LFM2 24B A2B at Q4_K_M (14.8 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.