Phi-3
Microsoft's small Phi-3 family. Mini (3.8B) is the best-in-class small chat; Medium (14B) competes with much larger models.
What Phi-3 is for
Phi-3 is a general-purpose open-weights model from Microsoft, released in 2024-04 with 2 sizes (Mini 3.8B, Medium 14B) and 4K (Mini), 128K (Medium) of context, licensed MIT. The catalogue records its strength as small chat, edge deployment.
Sizes and memory
Our catalogue does not carry a per-size memory breakdown for Phi-3. What it records instead is a floor of 3 GB and a recommended 6 GB for the default configuration. Until a per-size figure exists, treat those two numbers as the whole guidance rather than reading a size table that would have to be invented.
Hardware that runs Phi-3
Without a per-size memory figure there is no honest way to say which card holds Phi-3. Its recorded floor of 3 GB rules out every consumer card in our hardware guide, which tops out well below that.
What people use it for
The catalogue lists 4 recorded uses for Phi-3: chat, edge, mobile, low-resource. It is tagged chat, lightweight, small.
Where Phi-3 sits in the Phi family
Our catalogue holds 4 Phi entries. Ordered by release date, Phi-3 (2024-04) is the 1st of them, and Phi-4 Reasoning (2025-04) is newer. Phi-4 Mini carries the same recorded memory footprint, so choosing between it and this one is a question of behaviour rather than of hardware: Phi-3 is the one recorded here for small chat, edge deployment.
The MIT licence puts no usage ceiling on it, which matters if the thing you are building has users. That is the practical difference between this and a vendor community licence.
Context window: 4K (Mini), 128K (Medium)
The record gives the context window as 4K (Mini), 128K (Medium), which is not a single number we can convert into a word count. Check the model card before planning around a specific length.
How to run Phi-3 locally
Our catalogue does not record registry tags, so look the current tag up in the Ollama library or on Hugging Face before pasting these. The commands below are the shape of the workflow, not a copy-and-paste recipe.
Option 1: Ollama (simplest)
# find the tag for Phi-3 at ollama.com/library
ollama run <tag>Option 2: Mullama (production)
mullama pull <tag>
mullama run <tag>Option 3: llama.cpp (CLI)
# download a GGUF from Hugging Face, searching for "Phi-3 GGUF"
./llama-cli -m model.Q4_K_M.gguf -p "Hello, AI!"Option 4: Python with Mullama or llama-cpp-python
from mullama import Model, Context
model = Model.load("model.Q4_K_M.gguf", n_gpu_layers=99)
ctx = Context(model, n_ctx=4096)
print(ctx.generate("Hello, AI!", 256))Sources
Model record from src/data/models.json (2026-06-29); card capacities from src/data/gpus.json (2026-06-29). Upstream: Curated from ollama.com/library + community benchmarks + paperswithcode + huggingface.