Alibaba announces 'Qwen3.7-Plus,' an AI model comparable to Claude Opus 4.6.



Alibaba's AI research team, Qwen (Tongyi Lab), has announced its AI model, ' Qwen3.7-Plus .' Qwen3.7-Plus has outperformed the Claude Opus 4.6 in some benchmark tests.

Qwen3.7-Plus: Multimodal Agent Intelligence

https://qwen.ai/blog?id=qwen3.7-plus

Qwen3.7-Plus is a multimodal model that supports text, image, and video input. The following are the results of the agent performance benchmark test 'Terminal-Bench 2.0' performed on 'Qwen3.7-Plus', 'Qwen3.6-Plus', 'DeepSeek-V4-Pro', 'GLM-5.1', 'Kimi K2.6', 'Claude Opus 4.6', and 'GPT-5.4'. Qwen3.7-Plus recorded a score that exceeded that of the closed model Claude Opus 4.6.



The results of various benchmark tests are as follows. It can be seen that the Qwen3.7-Plus has performance comparable to the Claude Opus 4.6.



Qwen3.7-Plus is designed as a 'multimodal interactive hybrid agent' capable of operations such as 'recognizing and automatically operating an app's UI,' 'writing code based on image input,' and 'answering visual questions using knowledge from the internet.' The following example demonstrates the operation of 'using Python to obtain the differences between images and solve a spot-the-difference puzzle.'



The following post includes an embedded demo video titled 'Writing clone code for an existing app using Qwen3.7-Plus'.




Qwen3.7-Plus is available for use with the chat AI service Qwen Studio , as well as via Alibaba's AI. The price per million tokens is $0.40 (approximately 63.9 yen) for input, $0.08 (approximately 12.8 yen) for input with caching, and $1.60 (approximately 255.6 yen) for output. API details can be found at the following link.

ModelStudio Console
https://modelstudio.console.alibabacloud.com/ap-southeast-1?tab=doc#/doc/?type=model&url=2840914_2&modelId=qwen3.7-plus&serviceSite=international



Related Posts:

in AI, Posted by log1o_hf