Inkling-Small requirements — can you run it?
~263.7B parameters · 22 quantizations measured · updated 2026-07-31 · source: unsloth/Inkling-Small-GGUF
Every quantization: real file size vs what you actually need
| Quant | File (measured) | VRAM/RAM needed (8K ctx) | Runs on (first fit) |
|---|---|---|---|
| UD-IQ1_S | 74.8 GB | 91.2 GB | Mac 128GB unified |
| UD-IQ1_M | 78.8 GB | 96.1 GB | Mac 128GB unified |
| UD-IQ2_XXS | 82.3 GB | 100.2 GB | Mac 128GB unified |
| UD-IQ2_M | 82.4 GB | 100.4 GB | Mac 128GB unified |
| UD-Q2_K_XL | 87.9 GB | 107.0 GB | Mac 128GB unified |
| UD-IQ3_XXS | 97.9 GB | 119.0 GB | Mac 256GB unified |
| UD-IQ3_S | 107.5 GB | 130.5 GB | Mac 256GB unified |
| UD-Q3_K_M | 119.4 GB | 144.7 GB | Mac 256GB unified |
| UD-Q3_K_XL | 119.6 GB | 145.0 GB | Mac 256GB unified |
| UD-IQ4_XS | 127.4 GB | 154.4 GB | Mac 256GB unified |
| UD-IQ4_NL | 130.1 GB | 157.6 GB | Mac 256GB unified |
| UD-Q4_K_S | 152.3 GB | 184.3 GB | Mac 256GB unified |
| UD-Q4_K_M | 162.5 GB | 196.5 GB | Mac 256GB unified |
| UD-Q4_K_XL | 163.3 GB | 197.4 GB | Mac 256GB unified |
| UD-Q5_K_S | 184.2 GB | 222.6 GB | Mac 256GB unified |
| UD-Q5_K_M | 196.1 GB | 236.8 GB | Mac 512GB unified |
| UD-Q5_K_XL | 196.7 GB | 237.5 GB | Mac 512GB unified |
| UD-Q6_K | 218.9 GB | 264.2 GB | Mac 512GB unified |
| UD-Q6_K_XL | 239.7 GB | 289.2 GB | Mac 512GB unified |
| Q8_0 | 280.3 GB | 337.8 GB | Mac 512GB unified |
| UD-Q8_K_XL | 288.4 GB | 347.6 GB | Mac 512GB unified |
| BF16 | 527.4 GB | 634.4 GB | 1TB server RAM (CPU) |
KV cache uses a flat +20% context/overhead rule because this repo does not publish its config; exact numbers may differ for long contexts. MoE models keep all experts in memory — total size counts, not just active params.
Can I run Inkling-Small on my GPU? (UD-IQ1_M recommended quant)
| Hardware | Memory | Verdict |
|---|---|---|
| RTX 3060 12GB | 12 GB | no |
| RTX 4060 Ti 16GB | 16 GB | no |
| RTX 3090 24GB | 24 GB | no |
| RTX 4090 24GB | 24 GB | no |
| RTX 5090 32GB | 32 GB | no |
| RTX PRO 6000 96GB | 96 GB | no |
| Mac 16GB unified | 16 GB | no |
| Mac 32GB unified | 32 GB | no |
| Mac 64GB unified | 64 GB | no |
| Mac 128GB unified | 128 GB | RUNS |
| Mac 256GB unified | 256 GB | RUNS |
| Mac 512GB unified | 512 GB | RUNS |
| 32GB system RAM (CPU) | 32 GB | no |
| 64GB system RAM (CPU) | 64 GB | no |
| 128GB system RAM (CPU) | 128 GB | RUNS |
| 256GB system RAM (CPU) | 256 GB | RUNS |
| 1TB server RAM (CPU) | 1024 GB | RUNS |
| 2TB server RAM (CPU) | 2048 GB | RUNS |
How much VRAM does Inkling-Small need?
The honest answer: 96.1 GB at the UD-IQ1_M quant with 8K context — that is the measured file size (78.8 GB) plus KV cache and runtime overhead. The smallest published build needs 91.2 GB; the lossless one needs 634.4 GB.
Can I run it with CPU offload?
Yes, if combined RAM+VRAM ≥ file size — but generation speed is limited by memory bandwidth. Expect roughly: DDR5 dual channel ~10-30 tok/s for small MoE actives, single digits for big ones, 0.1-1 tok/s when experts page from disk. CPU-only is fine for batch jobs, painful for chat.
Why do other sites show different numbers?
Most pages compute weights from a formula (params × bits/8). We use the actual file sizes published in GGUF repos, which include the embedding table, unquantized tensors, and container overhead — that is why our numbers can differ from a naive calculation.
FAQ
How much VRAM does Inkling-Small need?
The measured answer: 96.1 GB at the UD-IQ1_M quant with 8K context. The smallest published build needs 91.2 GB; the lossless (F16/BF16) build needs 634.4 GB. These are real GGUF file sizes from Hugging Face, not formula estimates.
Can I run Inkling-Small on a 24GB GPU (RTX 3090/4090)?
Not at UD-IQ1_M (96.1 GB). Nothing fits 24GB — the smallest build needs 91.2 GB. Use the hosted API or a smaller model.
Can I run Inkling-Small on CPU with system RAM?
Yes, if your usable RAM is at least the file size. At UD-IQ1_M you need ~96.1 GB of RAM. CPU generation is memory-bandwidth-bound: expect single-digit tok/s for large MoE models, 10-30 tok/s for small active params, and below 1 tok/s when experts page from disk.
Why do VRAM numbers for Inkling-Small differ between sites?
Most sites compute weights from a formula (params x bits/8). ModelFit uses the actual GGUF file sizes published on Hugging Face, which include the embedding table, unquantized tensors, and container overhead — so our numbers reflect what you really download and load.