I have tested that Qwen3.6-Plus is significantly stronger than MiniMax M2.7 or multimodal, but this time there are not as many AI bloggers who have released rainbow fart reviews because Alibaba has run out of food for them. Probably because of the recent financial report issue, Alibaba cut its advertising budget. If I had to say, this is the right thing to do. It should have been conquered by its strength. When Qwen3.5 came out, it had so many budget preparations and was all over the place, but in reality, it turned out to be a mess. Alibaba's current strategy shift -Small model, small-sized Qwen3.5 fully open source, first time open source -The performance of the Qwen3.6-Plus model, which exceeds that of the MiniMax M2.7, is expected to delay open sourcing and follow the example of three other domestic models -Flagship model Qwen Max series, closed source The transition from Qwen3.5-Plus to Qwen3.5-Max takes about a month. After Qwen3.6-Plus is released, it is possible to regain the top spot in domestic models. However, during this period, DeepSeek V4 will also be released, and other domestic models may also need to be updated. Currently, it is 1-2 months behind schedule. It will not be difficult for Qwen to catch up after each company turns to closed source, because the biggest bug of domestic models (distillation) has been blocked by Claude. Distillation is catching up quickly at the beginning, and the intelligent improvement is very obvious. Later, it depends on the hard power of each company.