AI Potluck
Model components / Base / pretrained models

Phi-4

Microsoft

Microsoft's 14B-parameter small language model that punches well above its weight class on reasoning, math, and code, released December 2024 under MIT license. Trained on heavily curated and synthetic data, demonstrating the 'small models with high-quality data' thesis. As of May 2026 the Phi-4 line is still current: Microsoft has extended it with Phi-4-mini, Phi-4-multimodal, and Phi-4-reasoning / Phi-4-reasoning-plus (May 2025), plus Phi-4-reasoning-vision-15B (March 2026). No Phi-5 has been announced.

Phi-4, 14B dense model, weights released on Hugging Face under MIT (full release Jan 2025 following Dec 2024 announcement). Reasoning variants (Phi-4-reasoning / -reasoning-plus) released later under open weights but are separate SKUs; this record scores base Phi-4 14B.

Openness

3 high confidence
3.0
weights
open(MIT)
data
closed
code
closed
license
MIT(OSI)

MIT weights, which is genuinely permissive with no use restrictions, but neither the training corpus nor the training pipeline is published - Microsoft describes the synthetic data mix without releasing it. So 3/open_weights: the artifact is free to use and the recipe is not reproducible. Reaching 4 or 5 would need the data or the pipeline, not a more permissive license.

Adoption

4 high confidence
4.0

Phi-4 is a widely downloaded small model on Hugging Face, distributed via Azure AI Foundry, Ollama, and many local-inference stacks; download volume for the Phi-4 line is in the millions range. Level 4 reflects broad small-model adoption.

Capability

3 high confidence
3.0

14B model with outsized math/reasoning: ~56.1% on graduate-level (GPQA-style) science questions, 80.4% on competition math, strong AMC results; reportedly beats GPT-4 (its teacher) on some STEM. Strong-for-size but a small/older model vs 2026 frontier.

Unchanged since 2026-07-30 (last edited, not re-checked)