What GPU do I need to run microsoft/Phi-4-mini-reasoning?
3.8B parameters, published in BF16. View on Hugging Face
Set up Phi-4-mini-reasoningPhi-4-mini-reasoning is published by microsoft on Hugging Face, with 55,989 downloads and 240 likes to date. It's a Phi3ForCausalLM model built for text-generation, published natively in BF16.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision, plus a fixed overhead for KV-cache, activations, and fragmentation. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 7.1 GB | 8.6 GB | RTX 3060 | 1 | $0.110/hr |
| FP8 (quantized) | 3.6 GB | 4.3 GB | RTX 4070 | 1 | $0.121/hr |
| INT4 (quantized) | 1.8 GB | 2.1 GB | RTX 3070 | 1 | $0.088/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-4-mini-reasoning at its published (BF16) precision: 1× RTX 3060, at $0.110/hr per GPU ($0.110/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
Phi-4-mini-reasoning: common questions
Does Phi-4-mini-reasoning fit on a 12 GB GPU?
Yes. At BF16 it needs 8.6 GB of VRAM, so a 12 GB card holds it with 3.4 GB to spare. An 8 GB card is not enough for it at BF16.
What is the least VRAM Phi-4-mini-reasoning can run in?
2.1 GB, at INT4 (quantized), which fits a 6 GB card, against 8.6 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-4-mini-reasoning lower the GPU bill?
Yes. At BF16 the cheapest live fit is one RTX 3060 at $0.110/hr. At INT4 (quantized) it drops to one RTX 3070 at $0.088/hr, provided a quantized checkpoint exists for it.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More microsoft models
- Florence-2-base (232M, F16)
- phi-2 (2.8B, F16)
- VibeVoice-ASR (8.7B, BF16)
- Phi-3.5-vision-instruct (4.1B, BF16)
- Florence-2-large (777M, F16)
- phi-4 (14.7B, BF16)
Related reading: H100 pricing and specs, The best GPUs for AI, ranked, and Best GPU for LLM inference.
