DeepSeek Announces Significant API Price Increase
DeepSeek notified developers on August 6 that it intends to raise API prices by a significant, still unspecified amount in the near future. The notice provides neither a revised price schedule nor an effective date, according to TechNode and the South China Morning Post. Current V4 Flash rates are $0.14 per million input tokens and $0.28 per million output tokens, Bloomberg reports.
DeepSeek announced on August 6 that it intends to implement a significant increase in API pricing in the near future, without disclosing the new rates or an effective date. In a notice on its developer platform, the Hangzhou-based AI company advised users to plan their usage accordingly, according to the South China Morning Post.
TechNode likewise reported that the change remains prospective rather than effective and that DeepSeek has not published a new price schedule. Its current API billing differentiates input and output tokens as well as cached and uncached requests.
What is known about current pricing
Bloomberg reports that V4 Flash currently costs $0.14 per million input tokens and $0.28 per million output tokens. Those rates have been a central part of DeepSeek's appeal among developers evaluating model-serving costs.
For comparison, Bloomberg listed Moonshot's Kimi K3 at $3 per million input tokens and $15 per million output tokens, while Anthropic's Fable 5 was listed at $10 and $50, respectively. These figures are not directly interchangeable for all workloads: total API spend depends on input-output ratios, caching behavior, context length, and the model selected for a task.
Android Authority reported that the newly announced increase is separate from a mid-July change affecting peak-hour API usage. That earlier change doubled rates during certain time slots, according to the publication.
V4 Flash context
The pricing notice arrived roughly a week after DeepSeek released DeepSeek-V4-Flash-0731, according to the South China Morning Post. The outlet described the release as a 284-billion-parameter lightweight member of the V4 series, following an April preview of the series.
The public notice cited by the available reports does not specify the size of the increase or an effective date.
For teams using DeepSeek APIs in production, the immediate operational fact is uncertainty rather than a known new unit price. Companies managing comparable API pricing changes typically reassess workload-level token accounting, cache-hit rates, routing policies, and budget alerts once a provider publishes an effective date and rate card. Those controls matter because a change to either input or output pricing can affect applications differently, particularly retrieval-heavy systems and long-form generation workloads.
The announcement also narrows the usefulness of static provider cost comparisons. Practitioners comparing models on cost should retain the current DeepSeek rates as a historical baseline, then rerun estimates when the company releases its revised schedule.
Key Points
- 1DeepSeek has announced a significant upcoming API price increase, but published neither revised rates nor an effective date.
- 2Current V4 Flash pricing is $0.14 per million input tokens and $0.28 per million output tokens, Bloomberg reports.
- 3Comparable provider price changes commonly require teams to rerun workload-level cost models using token mix, caching, and routing assumptions.
Scoring Rationale
The announced increase affects a widely watched low-cost model API and can alter inference-cost assumptions for developers using DeepSeek in production. The impact remains bounded because neither the percentage increase nor the implementation date has been published.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
