MiMo v2 Flash — Hardware Requirements & GPU Compatibility
Chat
MiMo V2 Flash is Xiaomi's large-scale mixture-of-experts language model, built with nearly 310 billion total parameters. Designed for fast inference despite its size, the Flash variant prioritizes throughput and responsiveness, making it well-suited for interactive chat and real-time applications. Running it locally is a serious undertaking that demands high-end multi-GPU configurations, but it brings flagship-level Chinese and English language capabilities to users who have the hardware to support it.
337.1K downloads
650 likesSpecifications
- Publisher
- XiaomiMiMo
- Parameters
- 309.8B
- Architecture
- MiMoV2FlashForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 152,576
- Release Date
- 2026-02-27
- License
- MIT
Get Started
HuggingFace
How Much VRAM Does MiMo v2 Flash Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| IQ2_XXS | 2.20 | 85.6 GB | 98.4 GB | 85.19 GB | Importance-weighted 2-bit, extreme compression — significant quality loss |
| IQ2_XS | 2.40 | 93.3 GB | 106.1 GB | 92.94 GB | Importance-weighted 2-bit, extra small |
| IQ2_S | 2.50 | 97.2 GB | 110.0 GB | 96.81 GB | Importance-weighted 2-bit, small |
| IQ2_M | 2.70 | 105.0 GB | 117.7 GB | 104.55 GB | Importance-weighted 2-bit, medium |
| IQ3_XXS | 3.10 | 120.4 GB | 133.2 GB | 120.04 GB | Importance-weighted 3-bit |
| IQ3_XS | 3.30 | 128.2 GB | 141.0 GB | 127.79 GB | Importance-weighted 3-bit, extra small |
| Q2_K | 3.40 | 132.1 GB | 144.8 GB | 131.66 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 135.9 GB | 148.7 GB | 135.53 GB | 3-bit small quantization |
| IQ3_M | 3.60 | 139.8 GB | 152.6 GB | 139.40 GB | Importance-weighted 3-bit, medium |
| Q3_K_M | 3.90 | 151.4 GB | 164.2 GB | 151.02 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 155.3 GB | 168.1 GB | 154.89 GB | 4-bit legacy quantization |
| Q3_K_L | 4.10 | 159.2 GB | 171.9 GB | 158.76 GB | 3-bit large quantization |
| IQ4_XS | 4.30 | 166.9 GB | 179.7 GB | 166.51 GB | Importance-weighted 4-bit, compact |
| Q4_K_S | 4.50 | 174.7 GB | 187.4 GB | 174.25 GB | 4-bit small quantization |
| Q4_1 | 4.50 | 174.7 GB | 187.4 GB | 174.25 GB | 4-bit legacy quantization with offset |
| IQ4_NL | 4.50 | 174.7 GB | 187.4 GB | 174.25 GB | Importance-weighted 4-bit, non-linear |
| Q4_K_M | 4.80 | 186.3 GB | 199.1 GB | 185.87 GB | 4-bit medium quantization — most popular sweet spot |
| Q4_K_L | 4.90 | 190.1 GB | 202.9 GB | 189.74 GB | 4-bit large quantization |
| Q5_K_S | 5.50 | 213.4 GB | 226.2 GB | 212.98 GB | 5-bit small quantization |
| Q5_K_M | 5.70 | 221.1 GB | 233.9 GB | 220.72 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 256.0 GB | 268.8 GB | 255.57 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 310.2 GB | 323.0 GB | 309.79 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run MiMo v2 Flash?
Q4_K_M · 186.3 GBMiMo v2 Flash (Q4_K_M) requires 186.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 243+ GB is recommended. Using the full 262K context window can add up to 12.8 GB, bringing total usage to 199.1 GB. No single GPU has enough memory — multi-GPU or cluster setups are needed.
Which Devices Can Run MiMo v2 Flash?
Q4_K_M · 186.3 GB4 devices with unified memory can run MiMo v2 Flash, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Pro M2 Ultra (192 GB).
Runs great
— Plenty of headroom
Decent
— Enough memory, may be tightRelated Models
Frequently Asked Questions
- How much VRAM does MiMo v2 Flash need?
MiMo v2 Flash requires 186.3 GB of VRAM at Q4_K_M, or 310.2 GB at Q8_0. Full 262K context adds up to 12.8 GB (199.1 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 309.8B × 4.8 bits ÷ 8 = 185.9 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 13.2 GB (at full 262K context)
VRAM usage by quantization
Q4_K_M186.3 GBQ4_K_M + full context199.1 GB- Can NVIDIA GeForce RTX 5090 run MiMo v2 Flash?
No — MiMo v2 Flash requires at least 85.6 GB at IQ2_XXS, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
- What's the best quantization for MiMo v2 Flash?
For MiMo v2 Flash, Q4_K_M (186.3 GB) offers the best balance of quality and VRAM usage. Q4_K_L (190.1 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 85.6 GB.
VRAM requirement by quantization
IQ2_XXS85.6 GB~53%IQ3_XS128.2 GB~73%Q3_K_L159.2 GB~86%Q4_K_M ★186.3 GB~89%Q4_K_L190.1 GB~90%Q8_0310.2 GB~99%★ Recommended — best balance of quality and VRAM usage.
- Can I run MiMo v2 Flash on a Mac?
MiMo v2 Flash requires at least 85.6 GB at IQ2_XXS, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run MiMo v2 Flash locally?
Yes — MiMo v2 Flash can run locally on consumer hardware. At Q4_K_M quantization it needs 186.3 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is MiMo v2 Flash?
At Q4_K_M, MiMo v2 Flash can reach ~16 tok/s on AMD Instinct MI300X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: AMD Instinct MI300X → 5300 ÷ 186.3 × 0.55 = ~16 tok/s
Estimated speed at Q4_K_M (186.3 GB)
AMD Instinct MI300X~16 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of MiMo v2 Flash?
At Q4_K_M, the download is about 185.87 GB. The full-precision Q8_0 version is 309.79 GB. The smallest option (IQ2_XXS) is 85.19 GB.