What GPU do I need to run microsoft/Phi-mini-MoE-instruct?
7.6B parameters, published in BF16. View on Hugging Face
Set up Phi-mini-MoE-instructPhi-mini-MoE-instruct is published by microsoft on Hugging Face, with 64,032 downloads and 41 likes to date. It's a PhiMoEForCausalLM model built for text-generation, published natively in BF16.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision times a flat 1.2 overhead for activations and fragmentation. The KV-cache grows with context and is not in that factor; it is listed per model below where the architecture is published. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 14.2 GB | 17.1 GB | RTX A5000 | 1 | $0.176/hr |
| FP8 (quantized) | 7.1 GB | 8.5 GB | RTX 4070 Super | 1 | $0.121/hr |
| INT4 (quantized) | 3.6 GB | 4.3 GB | RTX 4070 Super | 1 | $0.121/hr |
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
Cheapest way to run Phi-mini-MoE-instruct at its published (BF16) precision: 1× RTX A5000, at $0.176/hr per GPU ($0.176/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
Phi-mini-MoE-instruct: KV cache by context length
The KV cache is the memory the attention layers hold for every token of context, on top of the weights. It is not part of the flat 1.2x overhead in the table above, grows with context length and with every concurrent request, and is why a long-context deployment needs more VRAM than the table shows.
KV-cache size for Phi-mini-MoE-instruct is not published for this architecture, so no figure is shown.
Phi-mini-MoE-instruct: common questions
Does Phi-mini-MoE-instruct fit on a 24 GB GPU?
Yes. At BF16 it needs 17.1 GB of VRAM, so a 24 GB card holds it with 6.9 GB to spare. A 16 GB card is not enough for it at BF16.
What is the least VRAM Phi-mini-MoE-instruct can run in?
4.3 GB, at INT4 (quantized), which fits a 6 GB card, against 17.1 GB at BF16. That row assumes an INT4 checkpoint has actually been published for this model, so check its Hugging Face page before planning around it.
Does quantizing Phi-mini-MoE-instruct lower the GPU bill?
Yes. At BF16 the cheapest live fit is one RTX A5000 at $0.176/hr. At FP8 (quantized) it drops to one RTX 4070 Super at $0.121/hr, provided a quantized checkpoint exists for it.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
Alternatives at this size
Other models for text-generation within about a third of Phi-mini-MoE-instruct's 7.6B parameters, from other model lines.
- Qwen3-8B (8.2B, BF16)
- Llama-3.1-8B-Instruct (8.0B, BF16)
- Mistral-7B-Instruct-v0.2 (7.2B, BF16)
- granite-4.1-8b (8.8B, BF16)
- DeepSeek-R1-0528-Qwen3-8B (8.2B, BF16)
More on Phi-mini-MoE-instruct
Fits on a 24 GB GPU at BF16: every model that fits in 24 GB.
Fits on a 12 GB GPU at FP8: every model that fits in 12 GB.
Fits on an 8 GB GPU at INT4: every model that fits in 8 GB.
Best chat and assistants models: how Phi-mini-MoE-instruct ranks against the rest.
Related reading: RTX A5000 pricing and specs, How much VRAM you need for LLMs, Serving LLMs with vLLM, vLLM vs TensorRT-LLM vs SGLang, and Best GPU for LLM inference.
