v2.2 · 2026-06-12
Phi-4 14B Instruct
phi-4-14b-instruct
Phi-4 14B Instruct is a 14B text chat model profile for local AI planning. This page records provider, license, context, quantization, hardware-fit estimates, setup hints, source links, and caveats so readers do not choose a model by name alone.
PROVIDER
Microsoft
FAMILY
Phi-4
MODEL TYPE
text-chat
PARAMETERS
14B
MODALITIES
text
ARCHITECTURE
dense decoder-only Transformer
CONTEXT WINDOW
16,384 tokens
TRAINING TOKENS
9.8T
RELEASE DATE
2024-12-12
GGUF REPOSITORIES
bartowski/phi-4-GGUF community conversion
Community llama.cpp GGUF quantization of the official Microsoft Phi-4 checkpoint.
SETUP HINTS
OLLAMA
Use ollama run phi4.LM STUDIO
Search for bartowski/phi-4-GGUF and select Q4_K_M or Q5_K_M based on memory.
LLAMA.CPP
Use hf.co/bartowski/phi-4-GGUF:Q4_K_M for a balanced local configuration.QUANTIZATION FILE SIZES (GGUF)
Q4_K_M
9.05 GB
Q5_K_M
10.6 GB
Q8_0
15.58 GB
Quantization note: Q4_K_M for default local use; Q5_K_M when 16GB+ RAM or 12GB+ VRAM is available.
RAM / VRAM ESTIMATES
MIN RAM
16 GB
COMFORTABLE RAM
32 GB
MIN VRAM
12 GB
COMFORTABLE VRAM
16 GB
These are conservative local-inference estimates, not official hardware requirements.
HARDWARE FIT
CPU only (no GPU)
slow
8 GB RAM
no
16 GB RAM
borderline
32 GB RAM
yes
8 GB VRAM
no
12 GB VRAM
borderline
24 GB VRAM
yes
Apple Silicon (unified memory)
borderline
Hardware fit values are conservative local-inference estimates based on GGUF size plus runtime and KV-cache overhead. Actual requirements depend on context length, quantization, runtime, GPU offload, and other running apps.
BEST FOR
·local reasoning and logic
·math-heavy prompts
·general chat
·medium-size workstation inference
AVOID IF
·8GB RAM or 8GB VRAM is the hard ceiling
·multimodal input is required
·non-English production use is required
CAVEATS
·Phi-4 is strong at reasoning but this is not the separate Phi-4-Reasoning model; reasoningTuned is therefore false.
·The model card states that Phi-4 is primarily English and not intended for multilingual use.
·GGUF files are community conversions, not official Microsoft GGUF releases.
·RAM and VRAM figures are estimates based on GGUF file size plus conservative runtime and KV-cache overhead, not official hardware requirements.
SOURCE URLS
FIELD EVIDENCE
modalitieshuggingface.co/microsoft/phi-4
architecturehuggingface.co/microsoft/phi-4
contextWindowTokenshuggingface.co/microsoft/phi-4
trainingTokenshuggingface.co/microsoft/phi-4
releaseDatehuggingface.co/microsoft/phi-4
hfGgufReposhuggingface.co/bartowski/phi-4-GGUF
q4FileSizeGbhuggingface.co/bartowski/phi-4-GGUF
q5FileSizeGbhuggingface.co/bartowski/phi-4-GGUF
q8FileSizeGbhuggingface.co/bartowski/phi-4-GGUF
ollamaollama.com/library/phi4
hardwareEstimateshuggingface.co/bartowski/phi-4-GGUF
LANGUAGES
English
CAPABILITIES
No special capabilities flagged.