Chinese researchers say they have created the world’s first brain inspired large language model, called SpikingBrain1.0. Unlike systems such as ChatGPT, which compare every word in a sentence to all others using high energy attention mechanisms, this model mimics the human brain’s selective firing of neurons. By focusing only on nearby words and recent context, it reduces energy consumption while still maintaining accuracy. The team from the Chinese Academy of Sciences claims SpikingBrain can run 25 to 100 times faster than traditional AI models, while also requiring less data for pre training. Importantly, it operates on China’s homegrown MetaX chip platform rather than relying on Nvidia GPUs, which have become harder to access due to US technology restrictions. According to the study, SpikingBrain consumes less than two percent of the training data compared to mainstream models yet still achieves competitive results. Scientists believe this approach could make AI development more efficient, less power hungry, and more widely deployable. While the research is still awaiting peer review, the claims highlight China’s growing efforts to develop independent AI technologies that bypass reliance on US hardware and potentially reshape the next generation of large scale models.