Can LFM3 VL 2B run on the Mac M4 Pro 24GB?
Yes. At 8K context the best fit is F16, using about 7.2 GB of the 16.1 GB usable on the Mac M4 Pro 24GB. Estimated speed ceiling: ~39 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|
| 2-bit (Q2_K) | 1.4 GB Fits | 1.4 GB Fits | 1.7 GB Fits | 380 |
| 3-bit (Q3_K_M) | 1.6 GB Fits | 1.6 GB Fits | 1.9 GB Fits | 291 |
| 4-bit (Q4_K_M) • | 1.7 GB Fits | 1.8 GB Fits | 2.0 GB Fits | 248 |
| 5-bit (Q5_K_M) • | 1.9 GB Fits | 1.9 GB Fits | 2.2 GB Fits | 217 |
| 6-bit (Q6_K) • | 2.0 GB Fits | 2.1 GB Fits | 2.4 GB Fits | 191 |
| 8-bit (Q8_0) • | 2.4 GB Fits | 2.5 GB Fits | 2.7 GB Fits | 148 |
| 16-bit (F16/BF16) • | 7.2 GB Fits | 7.2 GB Fits | 7.5 GB Fits | 39 |
• real GGUF file size from LiquidAI/LFM3-VL-2B-GGUF. Speed is a bandwidth ceiling (273 GB/s), not a benchmark. Methodology.
More
LFM3 VL 2B on every GPU and Mac · Every model that runs on the Mac M4 Pro 24GB · Apple Silicon guide · Try other settings in the calculator
Data: Hugging Face Hub (LiquidAI/LFM3-VL-2B), Apple specs. Data as of 2026-11-17.