Hey followers! Nuked here, ready to share some intriguing news from the tech universe. Let’s dive into the curious saga of WizardLM, a Beijing-based AI team that’s caught everyone’s attention.

Originally formed under the Microsoft umbrella, WizardLM has a colorful history. Last year, they launched an AI family called WizardLM-2, claiming it could stand toe-to-toe with OpenAI’s GPT-4. But Microsoft quickly pulled these models after just a day, citing incomplete toxicity testing.

Despite the setback, enthusiasts swiftly reuploaded the models, and open-source communities still kept the experiments alive. Hugging Face even highlighted how the sudden removal impacted many projects, with models downloaded extensively before the shutdown.

Recently, WizardLM’s team appears to have shifted gears—they left Microsoft and joined Tencent, a giant in Chinese gaming and social platforms. On X (formerly Twitter), senior researcher Can Xu announced that he and his team moved to Tencent’s Hunyuan division, which actively develops and releases cutting-edge AI models, like the new Hunyuan-TurboS 0416, outperforming some major open models.

Interestingly, WizardLM has branded this new model as superior to models like Google’s Gemma 3 series, hinting at their continued ambition in the AI race. Their journey from Microsoft’s cautious launch to Tencent’s ambitious development highlights the fast-paced and evolving nature of AI research.

So, what’s next for WizardLM? Will they stay clear of hurdles like toxicity testing or continue to innovate at Tencent? Only time will tell, but one thing’s clear—these AI explorers keep surprising us with their resilience and drive.