Qwen 3.7 Max Preview Lands: Rapid Dual‑Model Iteration Keeps China’s Lead in Text and Vision
The Qwen 3.7‑Max and Qwen 3.7‑Plus preview models debut with top‑15 global rankings in Arena, the only Chinese models in text and vision leaderboards, while a timeline analysis shows the Qwen series accelerating from 4‑6‑month releases to a 2‑3‑month cadence and introducing dense and MoE variants up to 235 B parameters.
Qwen 3.7 preview performance
Qwen 3.7‑Max‑Preview ranks 13th overall in the Arena text benchmark, making it the only Chinese model in the global top‑15 and raising Alibaba’s laboratory ranking to 6th.
Mathematics : 7th
Expert Prompt : 9th
Software/IT : 9th
Coding : 10th
In the visual benchmark, Qwen 3.7‑Plus‑Preview ranks 16th overall, lifting Alibaba’s visual ranking to 5th and being the sole Chinese model in the top‑10.
In the expert arena, Qwen 3.7‑Max‑Preview ranks 9th in the “expert prompt” track.
The only other Chinese model on the leaderboard is Xiaomi’s Mimo v2.5 Pro at 7th place.
Release cadence and model family
Early Qwen versions (2023‑2024) were released roughly every 4‑6 months. Starting with the Qwen 3 series, major versions appear every 2‑3 months.
The 2025 Qwen 3 family includes dense and Mixture‑of‑Experts (MoE) models ranging from 0.6 B to 235 B parameters. Two inference modes are defined:
Thinking – supports complex reasoning, long‑chain decisions, and agent tasks.
Non‑Thinking – optimized for low latency and fast response.
Preview releases are published first for community testing, followed by formal releases.
Recent acceleration
After the departure of former Qwen lead Lin Junyang (who posted “continue as planned” on X), the release pace has not slowed. In 2026, new versions have been released almost monthly, e.g., 3.5 → 3.6 → 3.7.
Reference URLs
https://x.com/Alibaba_Qwen/status/2056403591464984753
https://x.com/arena/status/2056400044862111757
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
DataFunTalk
Dedicated to sharing and discussing big data and AI technology applications, aiming to empower a million data scientists. Regularly hosts live tech talks and curates articles on big data, recommendation/search algorithms, advertising algorithms, NLP, intelligent risk control, autonomous driving, and machine learning/deep learning.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.