DeepSeek launching v4.1 flash cheaper and more capable than v4 pro
We are announcing the upcoming release of DeepSeek V4.1 Flash, scheduled for September 10, 2026. After extensive testing, we have found that V4.1 Flash surpasses V4 Pro in all key metrics, including performance, cost, speed, and task completion time. To ensure user satisfaction, we will route all Pro model requests to V4.1 Flash and bill at Flash's price until V4.1 Pro is released. We also plan to adjust Flash pricing, with off-peak rates at $0.003 for input cache hits, $0.15 for input cache misses, and $0.6 for output, and peak-hour rates double that. We invite users to provide feedback during comparative testing.
This is a significant upgrade. The fact that they are routing Pro requests to Flash and billing at Flash's price shows confidence in the new model. It will be interesting to see how the community responds to the price changes.
I'm curious about the actual performance benchmarks. The announcement says it surpasses V4 Pro in all key metrics, but I'd like to see independent testing and real-world results before switching.
The pricing strategy is smart. By offering lower off-peak rates, they can balance server load and reduce costs. This could make DeepSeek more competitive against other providers.