Group 1 - DeepSeek is developing an advanced AI model to compete with US rivals like OpenAI, aiming for a product launch in Q4 of this year [1] - The AI agent is seen as a key area following large models, featuring autonomous decision-making, task decomposition, and cross-application collaboration capabilities [1] - DeepSeek's R1 model, released in January, disrupted the global tech industry with its human-like reasoning capabilities [1] Group 2 - The new generation language model DeepSeek-V3.1 was officially launched on August 21, marking a step towards the agent era [1] - V3.1 incorporates UE8M0 FP8 Scale parameter precision and introduces three major changes: a hybrid reasoning architecture, improved thinking efficiency, and significantly enhanced agent capabilities [1]