DeepSeek Lists V4 Pro 0813 Build in Its API Catalog
DeepSeek's API pricing page now identifies the deepseek-v4-pro endpoint as DeepSeek-V4-Pro-0813. The page lists a 1 million-token context window, a 384,000-token maximum output, and prices of $0.435 per million uncached input tokens and $0.87 per million output tokens. The change is a dated build update to the V4 family, not a new model-family launch.
What changed
DeepSeek's current API pricing page identifies the model behind the deepseek-v4-pro endpoint as DeepSeek-V4-Pro-0813. Pandaily reported the catalog change on August 12 and described it as the Pro model moving into general availability. DeepSeek's own page confirms the build name, but it does not publish a separate release note or label the change as a new model-family launch.
The official page lists a 1 million-token context window and a 384,000-token maximum output for both V4 Pro and V4 Flash. It also says each endpoint supports thinking and non-thinking modes, JSON output, tool calls, the Responses API, an Anthropic-compatible API, prefix completion, and fill-in-the-middle completion in non-thinking mode.
Pricing and deployment implications
DeepSeek lists V4 Pro at $0.003625 per million cached input tokens, $0.435 per million uncached input tokens, and $0.87 per million output tokens. Those rates are a little more than three times the corresponding V4 Flash prices. The page also warns that DeepSeek plans a significant overall API price increase, without giving a date or final rate card.
For developers, the immediate operational point is version control. Applications that call the moving deepseek-v4-pro alias may now receive the 0813 build, so teams running evaluations should record the returned model version, repeat critical regression tests, and avoid attributing performance changes to the build without controlled comparisons.
The available official documentation does not disclose a change log, benchmark delta, training update, or exact rollout timestamp for 0813. Claims that the build closes a gap with frontier proprietary models therefore remain unverified by this release record. This update is best treated as an API snapshot change with documented serving limits and pricing, while the scale of any capability improvement remains open.
Key Points
- 1DeepSeek's API pricing page now maps deepseek-v4-pro to the dated DeepSeek-V4-Pro-0813 build.
- 2The documented endpoint supports a 1 million-token context window and up to 384,000 output tokens, with thinking and non-thinking modes.
- 3DeepSeek has not published a change log or independent performance evidence for the 0813 build, so teams should rerun their own regressions.
Scoring Rationale
The dated production build affects teams relying on DeepSeek's moving API alias and documents concrete context, output, and price limits. Impact is moderated because DeepSeek has not supplied a change log or verified capability delta for the 0813 snapshot.
Sources
Primary source and supporting 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
