Phi 3 Medium 4k Instruct — Hardware Requirements & GPU Compatibility
ChatCode
Phi 3 Medium 4k Instruct is a 14.0B-parameter open language model from Microsoft in the Phi 3 family. It supports a context window of up to 4,096 tokens. At Q4_K_M it needs about 9.10 GB of VRAM — see which GPUs and Macs can run it below.
Runs great on a 24 GB GPU
Runs from 8 GB (Q3_K_M · 4K)Best on 24 GB (Q4_K_M · ~72 tok/s · 4K)Macs from 16 GB (Q4_K_M · ~15 tok/s)
MIT
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Phi 3 Medium 4k Instruct
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve microsoft/Phi-3-medium-4k-instruct --max-model-len 4096vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3080 Ti (12 GB) (2 GB headroom) · Q4_K_M
- Generation speed
- ~65 tok/s
- generation speed
- Cost per 1M output tokens
- $0.52
- per 1M output tokens
How Much VRAM Does Phi 3 Medium 4k Instruct Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 6.7 GB | — | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 4K context · ~99 tok/s | |
| 7.5 GB | — | 3-bit medium — 3-bit medium quantization | Fits: 4K context · ~87 tok/s | |
| 9.1 GB | — | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 4K context · ~72 tok/s | |
| 10.7 GB | — | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 4K context · ~61 tok/s | |
| 12.2 GB | — | 6-bit, very good — 6-bit quantization, very good quality | Fits: 4K context · ~54 tok/s | |
| 14.7 GB | — | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 4K context · ~45 tok/s | |
| 28.6 GB | — | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 28.6 GB |
Selected Q4_K_M
4-bit medium quantization9.5GBat 4K context
- Weights
- 8.4 GB
- KV cache
- 0.8 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 14.5 GB free · ~72 tok/s
≥50 tok/s needs ≥699 GB/s · RTX 5090 ~128, RTX 3090 Ti ~72
Which Hardware Can Run Phi 3 Medium 4k Instruct?
Q4_K_M · 9.1 GBPhi 3 Medium 4k Instruct (Q4_K_M) requires 9.1 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 12+ GB is recommended. Using the full 4K context window can add up to 0.4 GB, bringing total usage to 9.5 GB. 40 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3080 Ti. 150 devices with unified memory can run Phi 3 Medium 4k Instruct, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Mini M2 Pro (16 GB).
12 GB
12 GB tier →fits · 2.5 GB free · up to 4K ctx
10 GPUs
RTX 3080 Ti~65Simulate output on RTX 3080 TiRTX 5070~48Simulate output on RTX 5070RTX 4070~36Simulate output on RTX 407020 GB
fits · 10.5 GB free · up to 4K ctx
1 GPU
RX 7900 XT~53Simulate output on RX 7900 XT32–40 GB
fits · 22.5 GB free · up to 4K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~128Simulate output on RTX 5090M5 Max36 GB~35Simulate output on Mac Studio M5 Max (36 GB)M532 GB~12Simulate output on MacBook Air 13" M5 (32 GB)Orin~15Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~10Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
How good is Phi 3 Medium 4k Instruct?
llmrun Score blends public benchmark boards into one number; 100 = the average open model.
Not enough data yet for an llmrun Score.
A Score needs results on enough independent boards. How the Composite Score is built
Benchmarks
Here are all 2 boards that have results for this model, sorted by their weight in the llmrun Score; badges show who ran the test.
- LMArena Text
LMArena · Oct 2026Arena
Arena1198.0
#147/188 open
- MMLU-Pro
TIGER-LabIndependent
Independent55.7%
#65/146 open
Alternatives to Phi 3 Medium 4k Instruct
Phi 3 family · 15 models →Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Apriel 1.6 15B ThinkerServiceNow-AI | — | 13.2 GB | ~67 | same size |
| Qwen1.5 14BAlibaba | — | 22.2 GB | ~63 | same size |
| Baichuan2 13B Basebaichuan-inc | — | 11.5 GB | ~76 | same size |
| Llama 2 13B Chat HFMeta | — | 11.5 GB | ~76 | same size |
| Vicuna 13B V1.3LMSYS | — | 8.6 GB | ~76 | same size |
| JiRackUltra 14BCMSManhattan | — | 12.4 GB | ~69 | same size |
Specifications
- Publisher
- Microsoft
- Family
- Phi 3
- Parameters
- 14.0B
- Architecture
- Phi3ForCausalLM
- Context Length
- 4,096 tokens
- Vocabulary Size
- 32,064
- Release Date
- 2024-05-07
- License
- MIT
Frequently Asked Questions
How much VRAM does Phi 3 Medium 4k Instruct need?
Phi 3 Medium 4k Instruct requires 9.1 GB of VRAM at Q4_K_M, or 28.6 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 14.0B × 4.8 bits ÷ 8 = 8.4 GB
KV Cache + Overhead ≈ 0.7 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 1.1 GB (at full 4K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 4090 run Phi 3 Medium 4k Instruct?
Yes, at Q8_0 (14.7 GB) or lower. Higher quantizations like BF16 (28.6 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
What's the best quantization for Phi 3 Medium 4k Instruct?
For Phi 3 Medium 4k Instruct, Q4_K_M (9.1 GB) offers the best balance of quality and VRAM usage. Q5_K_M (10.7 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 6.7 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Phi 3 Medium 4k Instruct on a Mac?
Yes — Mac Mini M2 Pro (16 GB) and 133 other Macs can run Phi 3 Medium 4k Instruct. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 9.1 GB of usable unified memory (RAM minus macOS overhead).
Can I run Phi 3 Medium 4k Instruct locally?
Yes — Phi 3 Medium 4k Instruct can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Phi 3 Medium 4k Instruct?
At Q4_K_M, Phi 3 Medium 4k Instruct can reach ~528 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~72 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 ÷ 9.1 × 0.65 = ~571 tok/s
Estimated speed at Q4_K_M (9.1 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Phi 3 Medium 4k Instruct?
At Q4_K_M, the download is about 8.38 GB. The full-precision BF16 version is 27.92 GB. The smallest option (Q2_K) is 5.93 GB.
Which GPUs can run Phi 3 Medium 4k Instruct?
40 consumer GPUs can run Phi 3 Medium 4k Instruct at Q4_K_M. Top options include AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 6900 XT, AMD Radeon RX 6700 XT. 26 GPUs have plenty of headroom for comfortable inference.
Which devices can run Phi 3 Medium 4k Instruct?
150 devices with unified memory can run Phi 3 Medium 4k Instruct at Q4_K_M (9.1 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (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.