Can LFM2 350M run on the RX 9070?
Yes. At 8K context the best fit is F16, using about 1.4 GB of the 15.2 GB usable on the RX 9070. Estimated speed ceiling: ~903 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|
| 2-bit (Q2_K) | 0.8 GB Fits | 0.9 GB Fits | 1.6 GB Fits | 4815 |
| 3-bit (Q3_K_M) | 0.8 GB Fits | 0.9 GB Fits | 1.7 GB Fits | 3694 |
| 4-bit (Q4_K_M) | 0.9 GB Fits | 1.0 GB Fits | 1.7 GB Fits | 2954 |
| 5-bit (Q5_K_M) | 0.9 GB Fits | 1.0 GB Fits | 1.8 GB Fits | 2534 |
| 6-bit (Q6_K) | 0.9 GB Fits | 1.0 GB Fits | 1.8 GB Fits | 2202 |
| 8-bit (Q8_0) | 1.0 GB Fits | 1.1 GB Fits | 1.9 GB Fits | 1699 |
| 16-bit (F16/BF16) | 1.3 GB Fits | 1.4 GB Fits | 2.2 GB Fits | 903 |
• real GGUF file size. Speed is a bandwidth ceiling (640 GB/s), not a benchmark. Methodology.
More
LFM2 350M on every GPU and Mac · Every model that runs on the RX 9070 · Other models that fit 16 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (LiquidAI/LFM2-350M), AMD specs. Data as of 2026-11-17.