[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

![ai_24x7 Avatar](https://lunarcrush.com/gi/w:24/cr:twitter::1708660994095058944.png) The AI Showdown 24x7 [@ai_24x7](/creator/twitter/ai_24x7) on x 2132 followers
Created: 2025-07-23 22:45:06 UTC

QWEN'S NEW AI MODEL DOMINATES BENCHMARKS

Alibaba Cloud has released Qwen3-235B-A22B-Instruct-2507, an enhanced version of its large language model, abandoning hybrid thinking modes in favor of separate Instruct and Thinking models for optimal quality. This update demonstrates superior performance across benchmarks like GPQA (77.5%), AIME25 (70.3%), and Arena-Hard v2 (79.2%), outperforming rivals such as Claude Opus X and DeepSeek-V3-0324. Available on Qwen Chat, Hugging Face, and ModelScope, it includes an FP8 version for efficiency.

![](https://pbs.twimg.com/media/Gwk6TT6W8AIqQ99.jpg)

XX engagements

![Engagements Line Chart](https://lunarcrush.com/gi/w:600/p:tweet::1948152420174794916/c:line.svg)

**Related Topics**
[superior](/topic/superior)
[hybrid](/topic/hybrid)
[$9988hk](/topic/$9988hk)
[alibaba](/topic/alibaba)
[coins ai](/topic/coins-ai)

[Post Link](https://x.com/ai_24x7/status/1948152420174794916)