Can I run MiMo V2 Flash locally?
MiMo V2 Flash by XiaomiMiMo has 310B parameters (16B active per token, 256 experts, 8 used per token). At 4-bit with an 8K context it needs about 176 GB of GPU memory. It fits on 6 of 65 GPUs and Macs we track.
Memory needed by quant and context length
GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from unsloth/MiMo-V2-Flash-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.
| Quant | Quality | 4K ctx | 8K ctx | 32K ctx | 128K ctx |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | rough | 106 GB | 106 GB | 110 GB | 123 GB |
| 3-bit (Q3_K_M) • | usable | 138 GB | 139 GB | 142 GB | 156 GB |
| 4-bit (Q4_K_M) • | good, most popular | 175 GB | 176 GB | 179 GB | 193 GB |
| 5-bit (Q5_K_M) • | very good | 205 GB | 206 GB | 209 GB | 223 GB |
| 6-bit (Q6_K) • | near lossless | 237 GB | 238 GB | 241 GB | 255 GB |
| 8-bit (Q8_0) • | effectively lossless | 307 GB | 308 GB | 311 GB | 324 GB |
| 16-bit (F16/BF16) • | original weights | 576 GB | 577 GB | 580 GB | 594 GB |
Which GPUs and Macs run MiMo V2 Flash?
Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).
Questions
How much VRAM does MiMo V2 Flash need?
About 176 GB at 4-bit (Q4_K_M) with an 8K context: 174 GB of weights (real GGUF file size), 1.1 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 308 GB.
What is the smallest GPU that runs MiMo V2 Flash?
No single consumer GPU in our table fits it at Q4_K_M with 8K context; the smallest option that does is the Mac M3 Ultra 512GB (512 GB).
Can I run MiMo V2 Flash on a Mac?
Yes. At Q4_K_M with 8K context it fits on a Mac M3 Ultra 512GB or larger, using the default macOS GPU memory limit.
Model facts
| Hugging Face repo | XiaomiMiMo/MiMo-V2-Flash |
| Parameters | 309,766,601,088 (310B) · ~16B active (computed from config) |
| Layers / hidden size | 48 / 4096 |
| Attention | GQA; 64 heads, 4 KV heads, head dim 192 |
| KV cache per token (FP16) | 144.0 KiB |
| Max context | 262,144 tokens |
| License | mit |
| Released on HF | Dec 16, 2025 |
| Downloads (30d) / likes | 16,245 / 756 |
Sources: Hugging Face model API · config.json (XiaomiMiMo/MiMo-V2-Flash) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.