DeepSeek API: models and pricing

Summary

DeepSeek's official price and limits page, checked on 2026-10-05. It lists two API models, both with 1M-token context and 384K maximum output, and peak and off-peak rates that differ by time of day. The publication date given is the last price change (2026-09-10).

Key points

• deepseek-flash (DeepSeek-V4.1-Flash), per 1M tokens, off-peak / peak: cache-hit input $0.003 / $0.006, cache-miss input $0.15 / $0.30, output $0.60 / $1.20. Supports vision. Concurrency limit 2,500.
• deepseek-v4-pro (DeepSeek-V4-Pro-0813): cache-hit input $0.022 / $0.044, cache-miss input $0.66 / $1.32, output $1.98 / $3.96. Text only. Concurrency limit 500.
• Off-peak rates are half the peak rates. Peak hours are 01:00–04:00 and 06:00–10:00 UTC, Monday to Friday, excluding Chinese public holidays.
• Both models support thinking mode, JSON output, tool calls, and an Anthropic-compatible API.

Why it matters

Cached input at fractions of a cent per million tokens, plus time-of-day pricing, shows that cache-aware and schedule-aware use now has a large effect on cost.