- New interface signals shift toward differentiated performance depending on task complexity
- Unreleased visual feature hints at broader multimodal push as competition intensifies for AI assistants
Chinese AI chatbot leader DeepSeek rolled out a major app update introducing tiered user modes on April 8, a move widely seen by industry observers as groundwork for the anticipated launch of its next-generation V4 model.
The updated interface adds two selectable options above the input bar — “Fast Mode” and “Expert Mode” — offering differentiated performance depending on task complexity.
The company has not formally announced a new model release, but analysts say the feature segmentation suggests a broader restructuring of capabilities spanning general chat, advanced reasoning and future multimodal interaction.
According to the app description, Fast Mode is designed for everyday conversations, prioritizing immediate responses and smooth interaction.
Meanwhile, Expert Mode targets complex queries and may involve brief wait times during peak usage, implying heavier computational workloads behind higher-level reasoning tasks.
The update marks the first time DeepSeek has introduced functional tiering since its debut in November 2023, signaling a shift toward more refined product positioning as competition heats up in China’s fast-moving AI assistant market.
A forthcoming ‘Vision’ mode?
A third option labeled “Vision” has also appeared in the interface but remains unavailable to most users, prompting speculation that visual understanding capabilities — potentially tied to a forthcoming model upgrade — are nearing release.
Earlier reporting by The Yangtzeer indicated that DeepSeek’s upcoming V4 model has been fully optimized to run on Ascend processors developed by Huawei Technologies.
This suggests a broader industry pivot toward domestically produced AI chips is underway as Chinese developers adapt to supply-chain constraints while expanding advanced AI services.
