Name: Mage-VL
Author: Microsoft Research

Mage-VL

Microsoft Research

Mage-VL is Microsoft's open-weight multimodal generative checkpoint for image and video understanding. It combines the Mage-ViT visual encoder, a two-layer projector, and the Qwen3-4B-Instruct-2507 language backbone. A cognition gate within the same checkpoint supports event-triggered streaming responses. The released BF16 weights use Apache 2.0; the language configuration supports 262,144 context positions. The exact complete checkpoint parameter count and release date are not established by the reviewed primary sources.

Benchmark scores

Benchmark	Score
Roboflow Vision Evals — Counting, Judged (September)	26.6
Roboflow Vision Evals — Counting, Strict (September)	26.6
Roboflow Vision Evals — Detection mAP@50 (September)	2.5
Roboflow Vision Evals — Detection mAP@50:95 (September)	0.7
Roboflow Vision Evals — Detection mAP@75 (September)	0.0
Roboflow Vision Evals — Extraction, Judged (September)	55.0
Roboflow Vision Evals — Extraction, Strict (September)	47.1
Roboflow Vision Evals — Identification, Judged (September)	59.4
Roboflow Vision Evals — Identification, Strict (September)	44.8
Roboflow Vision Evals — OCR (September)	60.6
Roboflow Vision Evals — Reasoning, Judged (September)	24.7
Roboflow Vision Evals — Reasoning, Strict (September)	18.8
Roboflow Vision Evals (September 2026 protocol)	38.1