Can I run MiMo V2.5 Pro locally?
MiMo V2.5 Pro by XiaomiMiMo has 1.02T parameters (44B active per token, 384 experts, 8 used per token). At 4-bit with an 8K context it needs about 590 GB of GPU memory. It fits on 1 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-V2.5-Pro-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 | 360 GB | 361 GB | 371 GB | 411 GB |
| 3-bit (Q3_K_M) • | usable | 430 GB | 432 GB | 442 GB | 481 GB |
| 4-bit (Q4_K_M) • | good, most popular | 589 GB | 590 GB | 600 GB | 640 GB |
| 5-bit (Q5_K_M) • | very good | 708 GB | 710 GB | 720 GB | 759 GB |
| 6-bit (Q6_K) • | near lossless | 791 GB | 792 GB | 802 GB | 842 GB |
| 8-bit (Q8_0) • | effectively lossless | 1015 GB | 1017 GB | 1027 GB | 1066 GB |
| 16-bit (F16/BF16) • | original weights | 1908 GB | 1910 GB | 1920 GB | 1959 GB |
Which GPUs and Macs run MiMo V2.5 Pro?
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).
Prefer an API?
Xiaomi: MiMo-V2.5-Pro is on OpenRouter at $0.435 per 1M input tokens and $0.87 per 1M output tokens. Price details and history →
Questions
How much VRAM does MiMo V2.5 Pro need?
About 590 GB at 4-bit (Q4_K_M) with an 8K context: 586 GB of weights (real GGUF file size), 3.3 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 1017 GB.
What is the smallest GPU that runs MiMo V2.5 Pro?
No single consumer GPU in our table fits it at Q4_K_M with 8K context.
Can I run MiMo V2.5 Pro on a Mac?
Not at Q4_K_M with 8K context on any Mac in our table, even a 512 GB Mac Studio, using the default GPU memory limit.
Model facts
| Hugging Face repo | XiaomiMiMo/MiMo-V2.5-Pro |
| Parameters | 1,023,182,841,600 (1.02T) · ~44B active (computed from config) |
| Layers / hidden size | 70 / 6144 |
| Attention | GQA; 128 heads, 8 KV heads, head dim 192 |
| KV cache per token (FP16) | 420.0 KiB |
| Max context | 1,048,576 tokens |
| License | mit |
| Released on HF | Apr 27, 2026 |
| Downloads (30d) / likes | 8,925 / 772 |
Sources: Hugging Face model API · config.json (XiaomiMiMo/MiMo-V2.5-Pro) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.