Can I run SingVL S1 4B on my PC?
New release: SingVL S1 4B was published on Hugging Face on Nov 11, 2026 by inclusionAI. Numbers update hourly as GGUF files appear. See everything new this week.
SingVL S1 4B (4.5B) needs about 3.6 GB of GPU memory at 4-bit with an 8K context. The cheapest consumer GPU in our table that runs it fully in VRAM at 4-bit or better is the Arc B580 (12 GB, $249 launch MSRP). The smallest Mac that runs it is the Mac M4 16GB.
SingVL S1 4B on every GPU and Mac
Best quant that fits fully in memory at 8K context with a speed ceiling; otherwise the llama.cpp offload plan (system RAM needed, estimated speed from DDR4-3200 to DDR5-6000). Macs and unified-memory PCs share one pool, so offload adds nothing there. How offload is estimated.
| Hardware | Memory | Verdict | Details |
|---|---|---|---|
| RX 7600 | 8 GB | Q8_0 · ~64 tok/s | details |
| RTX 4060 | 8 GB | Q8_0 · ~61 tok/s | details |
| RTX 5060 | 8 GB | Q8_0 · ~100 tok/s | details |
| RTX 3060 Ti | 8 GB | Q8_0 · ~100 tok/s | details |
| RTX 4060 Ti 8GB | 8 GB | Q8_0 · ~64 tok/s | details |
| RTX 3070 | 8 GB | Q8_0 · ~100 tok/s | details |
| RTX 3080 10GB | 10 GB | Q8_0 · ~170 tok/s | details |
| Arc B580 | 12 GB | F16 · ~54 tok/s | details |
| RTX 3060 12GB | 12 GB | F16 · ~43 tok/s | details |
| RX 7700 XT | 12 GB | F16 · ~51 tok/s | details |
| RTX 5070 | 12 GB | F16 · ~80 tok/s | details |
| RTX 4070 | 12 GB | F16 · ~60 tok/s | details |
| RTX 4070 Super | 12 GB | F16 · ~60 tok/s | details |
| RTX 4070 Ti | 12 GB | F16 · ~60 tok/s | details |
| RTX 3080 12GB | 12 GB | F16 · ~108 tok/s | details |
| Arc A770 16GB | 16 GB | F16 · ~66 tok/s | details |
| RTX 5060 Ti 16GB | 16 GB | F16 · ~53 tok/s | details |
| RTX 4060 Ti 16GB | 16 GB | F16 · ~34 tok/s | details |
| RX 7800 XT | 16 GB | F16 · ~74 tok/s | details |
| RX 7900 GRE | 16 GB | F16 · ~68 tok/s | details |
| RX 9070 | 16 GB | F16 · ~76 tok/s | details |
| RX 9070 XT | 16 GB | F16 · ~76 tok/s | details |
| RTX 5070 Ti | 16 GB | F16 · ~106 tok/s | details |
| RTX 4070 Ti Super | 16 GB | F16 · ~80 tok/s | details |
| RTX 4080 Super | 16 GB | F16 · ~87 tok/s | details |
| RTX 5080 | 16 GB | F16 · ~114 tok/s | details |
| RTX 4080 | 16 GB | F16 · ~85 tok/s | details |
| RX 7900 XT | 20 GB | F16 · ~95 tok/s | details |
| RX 7900 XTX | 24 GB | F16 · ~114 tok/s | details |
| RTX 3090 | 24 GB | F16 · ~111 tok/s | details |
| RTX 4090 | 24 GB | F16 · ~120 tok/s | details |
| RTX 3090 Ti | 24 GB | F16 · ~120 tok/s | details |
| RTX 5090 | 32 GB | F16 · ~213 tok/s | details |
| DGX Spark | 128 GB | F16 · ~32 tok/s | details |
| Ryzen AI Max+ 395 128GB | 128 GB | F16 · ~30 tok/s | details |
| Mac M1 16GB | 16 GB | F16 · ~8 tok/s | details |
| Mac M4 16GB | 16 GB | F16 · ~14 tok/s | details |
| Mac M2 24GB | 24 GB | F16 · ~12 tok/s | details |
| Mac M3 24GB | 24 GB | F16 · ~12 tok/s | details |
| Mac M4 Pro 24GB | 24 GB | F16 · ~32 tok/s | details |
| Mac M4 32GB | 32 GB | F16 · ~14 tok/s | details |
| Mac M1 Pro 32GB | 32 GB | F16 · ~24 tok/s | details |
| Mac M3 Pro 36GB | 36 GB | F16 · ~18 tok/s | details |
| Mac M4 Pro 48GB | 48 GB | F16 · ~32 tok/s | details |
| Mac M4 Pro 64GB | 64 GB | F16 · ~32 tok/s | details |
| Mac M1 Max 64GB | 64 GB | F16 · ~47 tok/s | details |
| Mac M4 Max 64GB | 64 GB | F16 · ~65 tok/s | details |
| Mac M2 Max 96GB | 96 GB | F16 · ~47 tok/s | details |
| Mac M3 Max 128GB | 128 GB | F16 · ~47 tok/s | details |
| Mac M4 Max 128GB | 128 GB | F16 · ~65 tok/s | details |
| Mac M2 Ultra 192GB | 192 GB | F16 · ~95 tok/s | details |
| Mac M3 Ultra 256GB | 256 GB | F16 · ~97 tok/s | details |
| Mac M3 Ultra 512GB | 512 GB | F16 · ~97 tok/s | details |
Questions
How much VRAM do I need for SingVL S1 4B?
About 3.6 GB at 4-bit (Q4_K_M) with an 8K context, including KV cache and runtime overhead. Longer contexts need more; the model page has the full table by quant and context.
Other popular checks: Can I run gpt-oss-120b?Can I run gpt-oss-20b?Can I run Llama 70B?Can I run Qwen3?Can I run DeepSeek R1? · Best GPU by budget · Mac vs PC
Model data: Hugging Face Hub API and GGUF file sizes. GPU specs: vendor pages. Speeds are estimates, not benchmarks. Methodology · Report an error.