DeepSeek V4.1 Flash

DeepSeek V4.1 Flash was released on 10 September 2026 and is served under the API alias `deepseek-flash`. DeepSeek's pricing documentation gives it a 1M-token context window and, more unusually, two prices for the same model depending on the hour: peak rates of $0.30 per million input tokens (cache miss) and $1.20 output, against off-peak rates of $0.15 and $0.60 — exactly half. Cached input falls to $0.006 peak and $0.003 off-peak. Peak hours are Monday to Friday, 01:00–04:00 and 06:00–10:00 UTC; everything outside that window bills at the lower rate. The same documentation confirms the model has replaced the earlier V4 Flash line: "the legacy names deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted, but the corresponding models have been retired, their requests are served by the DeepSeek-V4.1-Flash model and billed at the Flash price." The figure recorded on this board is the peak rate, because that is what a buyer pays during a normal working day. No benchmark results were published on a page that could be fetched, so none are recorded here.