DeepSeek API Pricing

Written 2 Sep 2026 · Prices verified 8 Oct 2026

Official rates, the off-peak discount, and what the same model costs on OpenRouter and Novita.

• DeepSeek's $12B Raise: The Funding Timeline and What It Means for API UsersDeepSeek's second 2026 round targets at least $12 billion (possibly $15B) with CATL and Tencent leading, on top of June's ~$7B first round — before a planned early-2027 IPO. The verified five-date timeline, where the money goes (self-built data centers, a 160,000-chip Huawei Ascend cluster, own inference silicon), and three practical reads for API users.
• ZeroClaw + DeepSeekDoes ZeroClaw work with DeepSeek's current API? The V4 reasoning_content fix timeline (#6059, closed June 2), the deepseek-flash model-name trap, and what the combo costs — verified against official issue trackers and DeepSeek's changelog.
• DeepSeek Context Caching ExplainedDeepSeek's disk cache is on by default: hits cost $0.003/M input vs $0.15/M missed on flash — 50x. Exact prefix-match rules, the three persistence points, how to read prompt_cache_hit_tokens, and what silently kills hits.
• Call DeepSeek V4 in 3 Minutes with the OpenAI SDKPort any OpenAI-SDK script to DeepSeek V4 in three minutes: two strings, a thinking mode that's on by default, and the model-ID trap most tutorials skip.
• Cheapest DeepSeek V4 API — 4 Options ComparedDeepSeek V4 Flash costs $9–$77 a month depending on where you run it. We compared all four options — and why most price quotes you'll find are wrong.
• DeepSeek Off-Peak Pricing ExplainedDeepSeek charges half price outside peak hours — peak is UTC 01:00–04:00 and 06:00–10:00, Mon–Fri. See those windows in your timezone and what it saves.
• DeepSeek V4 API Cost CalculatorCompare what DeepSeek V4 costs per month across OpenRouter, Novita AI and the official DeepSeek API, using your own token volume.