Simply put: DeepSeek has brought open-source models into the discussion range of top-tier closed-source models. It has three highlights: 1M super long context, significantly enhanced Agent capabilities, and further improved reasoning abilities. V4-Pro is already close to the level of Claude Opus 4.6 in non-thinking mode, but there’s still a gap compared to Opus 4.6 in thinking mode. V4-Flash, on the other hand, is a faster and cheaper version, suitable for large-scale API calls. Compared to top-tier closed-source models like GPT-5.5 and Claude Opus 4.6, DeepSeek-V4 may not be comprehensively stronger, but its significance lies in this: An open-source model is now approaching the core capability range of closed-source models.