Cursor internal bench: DeepSeek V4.1 Flash at $0.60–$0.98 per task with no higher scorer
The post says DeepSeek V4.1 Flash sits in Cursor's cheapest internal bench tier at $0.60–$0.98 per task with no higher-scoring model, set against roughly $480 billion wiped off Chinese AI stocks.
Reuters reported that Chinese AI-related stocks have already shed roughly $480 billion. The same week, DeepSeek V4.1 Flash showed up in Cursor as one of the lowest-cost options on their internal bench, sitting at $0.60–$0.98 per task with no higher-scoring model. The equities and the models are moving in different directions. Public market valuations of Chinese AI companies are being marked down while the open-weight models continue to land inside Western developer tools as the cheapest high-quality choice. That split is new. It also changes the competitive picture for labs that are still raising at high private valuations. The cost of strong open models keeps falling even as some public Chinese AI names are correcting. The gap between model availability and equity pricing is wider than it was six months ago.
This post quotes another post: x.com/1178773799287885825/status/211482658141155…