Phi Mini MoE Instruct — Hardware Requirements & GPU Compatibility
Chat
Phi-mini-MoE-instruct is Microsoft's lightweight Mixture-of-Experts model with 7.6 billion total and 2.4 billion active parameters. The card says it was compressed and distilled from the base model shared by Phi-3.5-MoE and GRIN-MoE using the SlimMoE approach, then post-trained with supervised fine-tuning and direct preference optimization for instruction following and safety. It is intended for English use in memory- and compute-constrained or latency-bound settings. Only 2.4 billion parameters are active per token, so it is fast, and the full weights fit on a single consumer GPU once quantized. The context length is 4,096 tokens, which is short by current standards. It is released under the MIT license, permitting commercial and research use, and was published in June 2025. It belongs to the SlimMoE series, which also includes the smaller Phi-tiny-MoE with 3.8 billion total and 1.1 billion active parameters.
32.8K downloads
42 likesSpecifications
- Publisher
- Microsoft
- Family
- Phi
- Parameters
- 7.6B
- Architecture
- PhiMoEForCausalLM
- Context Length
- 4,096 tokens
- Vocabulary Size
- 32,064
- Release Date
- 2025-06-23
- License
- MIT
Get Started
HuggingFace
How Much VRAM Does Phi Mini MoE Instruct Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 3.8 GB | 4.1 GB | 3.25 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 3.9 GB | 4.2 GB | 3.35 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 4.3 GB | 4.6 GB | 3.73 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 4.4 GB | 4.7 GB | 3.82 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 5.2 GB | 5.4 GB | 4.59 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 6.0 GB | 6.3 GB | 5.45 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 6.9 GB | 7.2 GB | 6.31 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 8.2 GB | 8.5 GB | 7.65 GB | 8-bit quantization, near-lossless |
est.= calculated VRAM estimate; no published GGUF file found for that quantization yet. Other rows are verified against real community uploads.
Which GPUs Can Run Phi Mini MoE Instruct?
Q4_K_M · 5.2 GBPhi Mini MoE Instruct (Q4_K_M) requires 5.2 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 7+ GB is recommended. Using the full 4K context window can add up to 0.3 GB, bringing total usage to 5.4 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroomWhich Devices Can Run Phi Mini MoE Instruct?
Q4_K_M · 5.2 GB58 devices with unified memory can run Phi Mini MoE Instruct, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Apple iPhone 17 Pro.
Runs great
— Plenty of headroom
Decent
— Enough memory, may be tightWhere to Download Phi Mini MoE Instruct
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Related Models
Frequently Asked Questions
- How much VRAM does Phi Mini MoE Instruct need?
Phi Mini MoE Instruct requires 5.2 GB of VRAM at Q4_K_M, or 15.9 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 7.6B × 4.8 bits ÷ 8 = 4.6 GB
KV Cache + Overhead ≈ 0.6 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 0.8 GB (at full 4K context)
VRAM usage by quantization
Q4_K_M5.2 GBQ4_K_M + full context5.4 GB- What's the best quantization for Phi Mini MoE Instruct?
For Phi Mini MoE Instruct, Q4_K_M (5.2 GB) offers the best balance of quality and VRAM usage. Q5_0 (5.3 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 2.7 GB.
VRAM requirement by quantization
IQ2_XXS2.7 GBIQ3_XS3.7 GBQ3_K_M4.3 GBQ4_K_M ★5.2 GBQ5_05.3 GBBF1615.9 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Phi Mini MoE Instruct on a Mac?
Phi Mini MoE Instruct requires at least 2.7 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 Phi Mini MoE Instruct locally?
Yes — Phi Mini MoE Instruct can run locally on consumer hardware. At Q4_K_M quantization it needs 5.2 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Phi Mini MoE Instruct?
At Q4_K_M, Phi Mini MoE Instruct can reach ~149 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~239 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 5.2 × 0.65 = ~484 tok/s
Estimated speed at Q4_K_M (5.2 GB)
~484 tok/s~239 tok/s~484 tok/s~442 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Phi Mini MoE Instruct?
At Q4_K_M, the download is about 4.59 GB. The full-precision BF16 version is 15.30 GB. The smallest option (IQ2_XXS) is 2.10 GB.
- Which GPUs can run Phi Mini MoE Instruct?
52 consumer GPUs can run Phi Mini MoE Instruct at Q4_K_M (5.2 GB). Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.
- Which devices can run Phi Mini MoE Instruct?
59 devices with unified memory can run Phi Mini MoE Instruct at Q4_K_M (5.2 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Apple iPhone 17 Pro, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.