models · microsoft

Phi-4-Mini

Microsoft Phi-4 compact model optimized for on-device and edge inference. 3.8B parameters with strong reasoning for its size.

Compare Verify

What this registry has checked

Hardened. We downloaded these weights and hashed them ourselves. The hash below is ours, not a publisher’s claim. It does not assert that the model is safe, accurate, or fit for any purpose. It says the bytes are the bytes.

Weight file SHA-256	aa3bc3998198d2d133021befe889079fb0337132e4516aed7d72be9e728b7490
Method	self-download-sha256

Re-check this yourself → Source → Full record →

What is recorded

Publisher	microsoft
Origin	United States
Licence	MIT
Parameters	3,836,021,760 (3.84B)
Context window	128K
Architecture	phi3
Layers	32
Hidden size	3072
Attention	GQA (24 heads, 8 KV)
Native precision	BF16
Pulls recorded here	580,018
Downloads reported upstream	401,935

Reported scores

MMLU-Pro	46.5
GPQA Diamond	33.1
AIME 2025	6.7
MATH-500	69.6
LiveCodeBench	12.6
IFBench	21.1
Humanity's Last Exam	4.4

Reported by a third party and recorded here, not re-run by us. Source: artificialanalysis.ai/api/v2.

Elsewhere in the registry

All models → Browse with filters → What will run on my hardware? →

Registry record as of 2026-06-13. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.