🇨🇳中国·AI 新闻·2026年9月9日·来源： Igeekphone China Phone, Tablet Pc, Vr, Rc Drone Ne

DeepSeek V4.1 Flash Set to Launch Around September 10, Reportedly Surpassing V4 Pro

内容仅提供英文版本

DeepSeek will launch its V4.1 Flash model around September 10, 2026, according to an official notice from the company’s open platform. The new model is expected to deliver improvements in performance, response speed, and operating costs compared to its predecessor, V4 Pro. DeepSeek states that V4.1 Flash has undergone extensive testing, showing it outperforms V4 Pro across key metrics. The company also plans to automatically route V4 Pro requests to V4.1 Flash upon launch, before the release of V4.1 Pro. DeepSeek is also adjusting its Flash series pricing, with new rates taking effect on September 10, 2026, including different peak and off-peak charges for input and output tokens.

DeepSeek’s upcoming V4.1 Flash model, set for a September 10, 2026 launch, represents a strategic move to balance AI performance with cost efficiency. The company claims V4.1 Flash surpasses V4 Pro in key metrics, including response speed and operating costs, and will automatically handle V4 Pro requests. This suggests DeepSeek views the Flash model as capable of taking on a broader range of demanding AI workloads, not just as a faster alternative. The introduction of native multimodal capabilities in V4.1 Flash, confirmed during its intermediate testing phase, is a crucial development. This feature could enable the model to process diverse data types more naturally, moving beyond text-centric tasks. For developers and businesses in Asia, the revised pricing structure, with lower off-peak rates for cached input at 0.02 yuan per million tokens, makes V4.1 Flash particularly appealing for large-volume AI applications where both speed and API costs are critical considerations. The shift points to DeepSeek’s confidence in V4.1 Flash’s ability to handle demanding applications while offering a more efficient option for real-world use. The company has not yet released full technical specifications, but the combination of a new architecture, multimodal support, and competitive pricing positions V4.1 Flash as a significant update within the Asian AI model landscape.