Correct result: \sqrt(3)/3. To test the geometry question on Qwen app and the thinking mode you can get the result: Thinking mode: correct answer \sqrt(3)/3. Without thinking mode: wrong answer. Overall, the 235B model is quite powerful compared to previous SOTA model. More about the key updates in Qwen3: Hybrid reasoning model, expanded language support (100+ languages), enhanced tool calling capabilities with Qwen-Agent supporting MCP. The newly open-sourced Qwen3 is China's first "hybrid reasoning model", a concept initially proposed by Claude3.7 and recently adopted by Gemini2.5 Flash. Essentially, this allows the model to toggle reasoning processes on/off. The primary purpose is to accelerate response generation for simple queries or time-sensitive scenarios by optionally disabling the thinking process while maintaining output quality. Previous approaches struggled to directly suppress reasoning steps in LLMs without retraining, as prompt engineering offered limited control. Qwen3 introduces two control methods: 1) A hard switch via enable_thinking parameter (True/False), and 2) When enabled, secondary soft switching through appending /no_think or /think tokens. Qwen also provides recommended parameter configurations to ensure optimal performance: Think mode: Temperature=0.6, TopP=0.95, TopK=20, MinP=0 Non-think mode: Temperature=0.7, TopP=0.8, TopK=20, MinP=0 Additionally, Qwen3 features specialized training for tool invocation, with Qwen-Agent now supporting MCP.

ChatGPT responded my question "What factors should I consider when choosing pacemaker for elderly people?" with 9 factors to consider when choosing pacemaker for elderly people, including Health Condition, Type of Pacemaker, Battery Life, Programming and Monitoring, Size and Comfort, Safety Features, Implantation Procedure, Hospital and Physician Experience, Cost and Insurance Coverage. In the end it advised me to consult with a cardiologist or cardiac electrophysiologist is essential to determine the most suitable pacemaker based on the specific medical condition and needs of the elderly person. The response are quite helpful and it mentions "Programming and Monitoring" and "Implantation Procedure", which are not covered by other AI agents.

I asked ChatGPT a question about "Requirements of selling a house in California". And ChatGPT lists 10 required documents of selling houses in California, including Seller's Disclosures, Transfer Disclosure Statement (TDS), Natural Hazard Disclosure Statement (NHDS), Local Ordinances, Title Transfer and Escrow, Tax and Fee Payments, Closing Documents, Compliance with Fair Housing Laws, Real Estate Agent Requirements, Capital Gains Tax. I compared the results with responses from other AI, and it seems like "Megan's Law Disclosure", "Purchase Agreement", "Deed" are missing. And ChatGPT provides unique answers like Fair Housing Laws, etc. Besides, additional links of trustworthy information resources are not provided. So I can only give it an average rating. Prompt: You are a lawyer specialized in real estate law. I will consult you a few questions. And my questions include "What are the legal requirements for selling a house in California?"

I asked ChatGPT to act as an insurance agent, who will answer questions about insurance, including "Should I buy health insurance for my mom who is above 50 years old already?" and "My mom has broken her legs and is on wheels right now. And I am wondering if it's economic to buy additional health insurance". ChatGPT replied with very general and vague response like considering "Medical Expenses", "Coverage Benefits" and "Financial Protection". And for my second question about my mom's situation, broken legs on wheel chair, it still gave vague response like "Current Needs", "Insurance Coverage", "Cost-Benefit Analysis", "Consultation". So not very helpful...