Can SmolLM2 360M run on the RTX 4070?
Yes. At 8K context the best fit is F16, using about 1.5 GB of the 11.4 GB usable on the RTX 4070. Estimated speed ceiling: ~695 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | Max tok/s (est., 8K) |
|---|---|---|---|
| 2-bit (Q2_K) • | 0.9 GB Fits | 1.0 GB Fits | 2305 |
| 3-bit (Q3_K_M) • | 0.9 GB Fits | 1.0 GB Fits | 2148 |
| 4-bit (Q4_K_M) • | 0.9 GB Fits | 1.1 GB Fits | 1863 |
| 5-bit (Q5_K_M) • | 0.9 GB Fits | 1.1 GB Fits | 1738 |
| 6-bit (Q6_K) • | 1.0 GB Fits | 1.1 GB Fits | 1372 |
| 8-bit (Q8_0) • | 1.0 GB Fits | 1.2 GB Fits | 1304 |
| 16-bit (F16/BF16) • | 1.3 GB Fits | 1.5 GB Fits | 695 |
• real GGUF file size from mradermacher/SmolLM2-360M-GGUF. Speed is a bandwidth ceiling (504 GB/s), not a benchmark. Methodology.
More
SmolLM2 360M on every GPU and Mac · Every model that runs on the RTX 4070 · Other models that fit 12 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (HuggingFaceTB/SmolLM2-360M), NVIDIA specs. Data as of 2026-11-17.