Skip to content
NO FEED

Understand the changes. Choose the tools. Put AI to work.

← News
News

DeepSeek peak/off-peak API rates are live as of 16 Aug 16:00 UTC.

Peak hours are 01:00–04:00 and 06:00–10:00 UTC; all other hours are off-peak at half of peak.1 Flash peak cache-miss input $0.44 / 1M and output $1.32 / 1M (off-peak $0.22 / $0.66). Pro peak $1.32 / $3.96 (off-peak $0.66 / $1.98). Cache-hit input is a separate, much lower tier.1

Issue 009 printed the notice before the cutover. The cutover held. Model versions on the same page: Flash-0731, Pro-0813.1

Breaking

Meter DeepSeek by UTC hour. Prefer off-peak for batch and long loops; pin Pro-0813 or Flash-0731 so a silent version cell cannot surprise you.

Deeper — hours and pins
  • Peak: 01:00–04:00 and 06:00–10:00 UTC only.1
  • Concurrency: Flash 2500, Pro 500.1
  • OpenAI- and Anthropic-format base URLs unchanged; Responses API and tool calls still listed.1

Sources

  1. DeepSeek docs — Models & Pricing (peak/off-peak live) · read
Originally part of the 17 August 2026 edition.

The useful part, in your inbox.

One complete edition every weekday, with practical takeaways.