Arena.ai
@arena
In the Vision Arena, Qwen3.7 Plus Preview makes @Alibaba_Qwen the #5 lab, ranking #16 overall. https://t.co/BppubI1h2B
6retweets70likes
View on XAlibaba Qwen3.7 Preview Enters Arena Top 15 for Text and Vision
Arena.ai, a community-driven platform that ranks AI models through blind human preference testing, added Alibaba’s
Qwen3.7 Preview models to its global leaderboards. The Qwen3.7 Max Preview variant secured the #13 spot in the Text Arena, while Qwen3.7 Plus Preview entered the Vision Arena at #16.- Text Arena rank
- #13 overall
- Vision Arena rank
- #16 overall
- Math rank
- #7
- Expert prompt rank
- #9
- Software and IT rank
- #9
- Coding rank
- #10
The update highlights Alibaba’s rapid development cycle, following the recent Qwen 3.5 Max Preview expert rankings and Qwen3.6 Plus coding performance. Qwen3.7 shows particular strength in specialized domains, ranking #7 in math and #9 for expert-level prompts. This suggests the model family is moving beyond general conversation into complex reasoning.
You can now evaluate these preview models across multiple modalities on the Arena platform to compare their performance against other top-tier systems. The high rankings in software and IT (#9) and coding (#10) make these models strong candidates for technical workflows. Detailed leaderboard data is available through the Arena web interface.
Still wondering? A few quick answers below.
Alibaba Qwen3.7 Preview models have achieved high rankings on the Arena.ai leaderboards through blind human testing. The Qwen3.7 Max Preview model is currently ranked #13 overall in the Text Arena, while the Qwen3.7 Plus Preview model holds the #16 spot in the Vision Arena for multimodal tasks.
Qwen3.7 Max Preview shows significant strength in technical and reasoning domains. On the Arena leaderboards, it ranks #7 for math-specific prompts and #10 for coding tasks. It also holds the #9 spot in both the expert-only prompt category and the software and IT category, indicating high proficiency in complex technical reasoning.
The Qwen3.7 Preview results place Alibaba among the top six global AI labs on the Arena.ai leaderboards. Arena.ai measures model performance using blind human preference testing across specialized categories like math, coding, and vision, making the rankings a reflection of real-world usability rather than just benchmark scores.
Both are preview variants of Alibaba's Qwen3.7 model family currently listed on Arena.ai. The Max variant currently achieves a higher rank in the Text Arena at #13, while the Plus variant is specifically evaluated in the Vision Arena at #16. The distinction between Max and Plus typically refers to model size and capability tiers within the Qwen family, though specific parameter counts are not yet disclosed for preview models.
