Phi-4-Reasoning-Vision-15B

Phi-4-Reasoning-Vision-15B.

Microsoft Released 4 Mar 2026 Family Phi-4 The lab's model card

Benchmarks with a score

Ranked first, out of five or more models

In the top three

Microsoft Mar 2026

Open weights Small

No Epoch Capabilities Index score yet

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Benchmark	Score		Rank	Source
HallusionBench	64.4%		#10 of 23	Model card

Benchmark	Score		Rank	Source
ScreenSpot-v2	88.2%		#12 of 14	Model card

Benchmark	Score	Rank	Source
MathVista	75.2%	#30 of 64	Model card
ChartQA	83.3%	#23 of 37	Model card
AI2D	84.8%	#34 of 44	Model card
MMStar	64.5%	#26 of 30	Model card
MMMU	54.3%	#88 of 97	Model card
OCRBench	76%	#27 of 28	Model card
MATH-Vision (MathVision)	36.2%	#41 of 42	Model card
MathVerse	44.9%	#12 of 12	Model card

Benchmark	Score		Rank	Source
Harmful content defect rate (Azure AI Evaluation SDK, Microsoft)	1.40%		#1 of 3	Model card

The benchmarks the lab chose to headline, with the scores it gave.

Every result in the longer documents the lab published with it. 1 of them isn't in the launch post.