Qwen QwQ 32B-Preview a New AI Reasoning Model 🧠 #llm #ai #ainews #openai #opensourceai #gpt #qwen #aimodels #artificialintelligence #genai

Alright, a new open source AI model just dropped and it's already making waves in reasoning and math benchmarks. It's called QWQ32B Preview by the QWEN team. It stands for QWEN with questions, and it's pronounced like quill. This model competes directly with OpenAI's O1 Preview model, but it's fully open source and you can run it on your computer. This model is insane. For reasoning tasks, it scored 65.2% on GPQA, 50% on AIME, and a whopping 90.6% on Math500. It's also crushing it on LiveCodeBench for programming tasks too. It is 32 billion parameters, which is quite small and has a 32k context length. Using the test, please add a pair of parentheses to the incorrect equation to make the equation true. Here's the entire thought process of solving the problem. It is amazing how fast we're getting reasoning models at the level of O1 Preview. We had DeepSeq R1 Lite previewed last week, now we have QWQ. If you want to try it out, they have a free preview on Hugging Face Spaces here. You can also download the model and run it yourself. Like and follow for more AI.