Can't Wait for the Final Version? Qwen3.7 Preview Model is Now Available for Experience! Mathematics/Programming/Multimodal Capabilities Have Been Significantly Enhanced, Get a First Look at the Thinking Mode

Large model iterations have once again accelerated. Two preview models of the Alibaba Tongyi Qwen 3.7 series—**Qwen3.7-Max-Preview** and **Qwen3.7-Plus-Preview**—have quietly been launched on Qwen Chat and Arena AI (formerly LMArena), and are likely to be officially announced at the Alibaba Cloud Summit on May 20th.

The two preview versions released this time have clear positioning:

Qwen3.7-Max-Preview: A flagship "brain" preview, focusing on the top comprehensive performance, currently only opening the thinking mode, with tool capabilities such as search and code interpreter not yet unlocked;

Qwen3.7-Plus-Preview: A high-performance balanced version preview, also focusing on reasoning and logical expression, with the toolchain to be gradually opened in the future.

Although still in "preview" mode, the actual performance has already attracted attention. In the text capability ranking of Arena AI, Qwen3.7-Max-Preview ranks 13th overall, helping Alibaba's Tongyi Qwen to rank sixth globally among similar labs; in specific fields, it is even more impressive: 7th in mathematical reasoning, 9th in expert-level applications, 9th in software and IT fields, and 10th in programming ability, demonstrating solid multi-task generalization capabilities.

In terms of visual capabilities, Qwen3.7-Plus-Preview ranks 16th overall, helping Tongyi Qwen to reach 5th among all labs in the visual ranking; in the "Expert Arena," which is only open to expert-level prompts, Qwen3.7-Max-Preview also secured the 9th position, indicating that its stability in high-difficulty and high-precision tasks is rapidly approaching the top tier.

Monthly AI Computing Power Bill! Major Telecommunications Operators All Enter the Market and Launch Token Plans

During World Telecommunication Day, the three major telecom operators packaged AI large model tokens as monthly subscription products, supporting phone bill payment and cross-platform use. This allows ordinary users, developers, and SMEs to access computing resources as easily as buying data packages, breaking the exclusive billing barriers of internet giants and making large model basic energy accessible to the public.....

No More Parameter Wars, Focus on Practical Application! Sohu Clearly States Its AI Strategy of Rational Application: How Mid-sized Tech Companies Can Forge a Differentiated Path Between Giants?

At the 2026 Tech Forum, Sohu CEO Charles Zhang outlined an AI strategy: avoid chasing the 'hundred-billion parameter' race, focus on practical, profitable applications. Sohu opts to be the smartest user of models, not their creator, investing in efficiency improvements to steer clear of prestige-driven competition.....

Tongyi Qianwen Makes a Major Upgrade: the New Flagship Qwen3.7 Preview Version Debuts Globally

Alibaba Cloud's Tongyi Qianwen large model has launched preview versions Qwen3.7-Max-Preview and Qwen3.7-Plus-Preview, available on the large model arena and official Qwen Chat. These next-generation flagship previews exclusively support a 'thinking mode,' focusing on high-difficulty logical reasoning and deep computation, with an official release expected at the Alibaba Cloud Summit.....

Baidu Releases Q1 2026 Earnings Report: Total Revenue of 32.1 Billion Yuan, AI Business Revenue Share Exceeds 52% for the First Time

Baidu's Q1 2026 report shows total revenue of 32.1 billion yuan, with general business revenue of 26 billion yuan, up 2% year-on-year. AI business revenue reached 13.6 billion yuan, surpassing 52% of general business for the first time, becoming a core growth engine. Executives discussed large model iteration, computing power layout, and commercialization paths during the earnings call.....

China's Large Model Weekly Query Volume Exceeds the US for Three Consecutive Weeks, Tencent Hy3 Preview Weekly Query Volume Reaches 2.66 Trillion

Global AI competition intensifies as latest data shows China's large model weekly call volume surpasses the US for three consecutive weeks. From May 11 to 17, global AI model total calls reached 26.9 trillion tokens, up 4.7% week-on-week. China contributed 7.693 trillion tokens, while the US had 4.24 trillion, with China clearly leading.....