DeepSeek V4 Pro official API update: Multiple test performances approach Fable 5

Odaily News: DeepSeek has officially upgraded V4 Pro from preview to general availability. On the official pricing page, the model version has been updated to DeepSeek-V4-Pro-0813. According to the benchmark comparison chart released in the official community, the DeepSeek V4 Pro official version (DeepSeek-V4-Pro-0813) approaches Fable 5 levels across multiple tests, showing a significant improvement over the previous preview version. DeepSeek V4 Pro is priced notably higher than V4-Flash. Per 1M tokens, cache-hit input costs 0.025 RMB, cache-miss input costs 3 RMB, and output costs 6 RMB. In comparison, V4-Flash is priced at 0.02 RMB, 1 RMB, and 2 RMB respectively. (Jin Shi)