Name: Qwen3.7 Flash
Author: Alibaba

Qwen3.7 Flash

Alibaba

Alibaba Cloud's hosted Qwen3.7 Flash vision-language model. The official catalog documents text, image and video input, text output, a million-token context, tool calling and optional thinking. A precise public release date is not established by the catalog.

Benchmark scores

Benchmark	Score
Capability	145.4
Chess Puzzles	23.0
Epoch Capabilities Index (ECI)	145
FrontierMath Tiers 1-3 (v2) - Private	19.3
GPQA Diamond	82.3
OTIS Mock AIME 2024-2025	86.7
Roboflow Vision Evals — Counting, Judged (September)	46.0
Roboflow Vision Evals — Counting, Strict (September)	46.0
Roboflow Vision Evals — Detection mAP@50 (September)	42.8
Roboflow Vision Evals — Detection mAP@50:95 (September)	31.1
Roboflow Vision Evals — Detection mAP@75 (September)	32.9
Roboflow Vision Evals — Extraction, Judged (September)	77.3
Roboflow Vision Evals — Extraction, Strict (September)	78.3
Roboflow Vision Evals — Identification, Judged (September)	84.4
Roboflow Vision Evals — Identification, Strict (September)	84.4
Roboflow Vision Evals — OCR (September)	84.1
Roboflow Vision Evals — Reasoning, Judged (September)	61.6
Roboflow Vision Evals — Reasoning, Strict (September)	60.9
Roboflow Vision Evals (September 2026 protocol)	61.5