DeepSeek launching v4.1 flash cheaper and more capable than v4 pro

DeepSeek is set to launch its V4.1 Flash model on September 10, 2026, according to official announcements. Extensive testing by DeepSeek has shown that the V4.1 Flash model outperforms the V4 Pro across several key metrics such as performance, cost, speed, and task completion time. Upon the official launch of the V4.1 Flash model, all requests to the V4 Pro model will automatically be sent to the V4.1 Flash model and will be billed at the Flash model's price.

If users experience any issues while comparing V4 Pro and V4.1 Flash, they are encouraged to provide feedback.

The pricing for the Flash series will be adjusted starting at 12:00 Beijing Time on September 10, 2026. During off-peak hours, the cost per input cache hit is $0.003, while input cache misses and outputs will cost $0.15 and $0.6 respectively. During peak hours, these prices will be double the off-peak rates. Users are advised to plan their usage accordingly. More information can be found at the provided comments URL.