ME News reports that on April 24 (UTC+8), according to monitoring by Beating, DeepSeek V4 API has officially launched V4-Pro and V4-Flash, with pricing and compute planning disclosed via the official WeChat public account. V4-Flash directly replaces V3.2 (deepseek-chat), without any price increase—in fact, prices have decreased: cached input remains unchanged at 0.2 RMB per million tokens; uncached input has dropped from 2 RMB to 1 RMB (a 50% reduction); output has fallen from 3 RMB to 2 RMB (a 33% reduction). Context length has been expanded from 128K to 1M, effectively providing eight times the context at a lower cost. The legacy model names deepseek-chat and deepseek-reasoner will be discontinued on July 24, 2026; they currently point to the non-reasoning and reasoning modes of V4-Flash, respectively. V4-Pro is a new premium tier: cached input at 1 RMB, uncached input at 12 RMB, and output at 24 RMB per million tokens—output pricing is eight times that of V3.2. DeepSeek notes in its pricing table that due to limited high-end compute capacity, Pro’s service throughput is currently very restricted, and prices are expected to drop significantly after the bulk release of Ascend 950 super nodes later this year. Both models support non-reasoning and reasoning modes, with the reasoning mode offering two intensity levels via the reasoning_effort parameter: high and max. DeepSeek stated in its announcement, “Starting now, 1M context will be standard across all official DeepSeek services.” (Source: BlockBeats)
DeepSeek V4 API Launches Flash and Pro Models with Price Cuts and 8x Context Expansion
KuCoinFlashShare
Summary
On April 24 (UTC+8), the DeepSeek V4 API launched the V4-Pro and V4-Flash models, featuring updated pricing and expanded context capacity. V4-Flash replaces V3.2 with a 50% reduction in input cost and a 33% reduction in output cost, while context length increases from 128K to 1M tokens. V4-Pro, a new top-tier model, has output pricing 8 times higher than V3.2. DeepSeek anticipates prices will decline in H2 with the introduction of Ascend 950 nodes. Crypto price movements and market sentiment, as reflected in the Fear & Greed Index, may influence adoption of these new models.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.