DeepSeek's search and fetch calls stop being free on 1 November
DeepSeek's pricing page, read on 26 October 2026, says server-side search and fetch calls made through DeepSeek Harness, Claude Code or Codex are free today and will be billed at $0.75 per 1,000 successful requests from 00:00 Beijing time on 1 November 2026. The same page says off-peak rates are half the peak rates, and that two older model names are retired.
In this article
What DeepSeek's page says
Everything in this section is a statement on DeepSeek's own pricing page, read on 26 October 2026. The page is JavaScript-rendered, so we read it in a browser and saved the rendered text with a hash.
Each of Search and Fetch is billed per 1,000 successful requests. DeepSeek says server-side calls made through DeepSeek Harness, Claude Code or Codex are free now and billing starts at 00:00 Beijing time on 1 November 2026.
Off-peak rates are half the peak rates. Peak is 01:00 to 04:00 and 06:00 to 10:00 UTC on weekdays, about 20.8% of the hours in a week.
DeepSeek says deepseek-v4-flash and deepseek-v4-flash-vision-exp are retired. Requests under those names are served by DeepSeek-V4.1-Flash-1017 and billed at the Flash price. Use deepseek-flash.
Both models list a 1M context length and a maximum output of 384K tokens.
When prices double
DeepSeek prices the same tokens differently by hour. On a weekday, 7 of 24 hours are peak.
Hours of a weekday in UTC. Shaded hours are peak. DeepSeek says all weekend hours and Chinese public holidays are off-peak. Convert to your own time zone before relying on this.
| Per million tokens | Flash off-peak | Flash peak | Pro off-peak | Pro peak |
|---|---|---|---|---|
| Input, cache hit | $0.003 | $0.006 | $0.022 | $0.044 |
| Input, cache miss | $0.15 | $0.30 | $0.66 | $1.32 |
| Output | $0.60 | $1.20 | $1.98 | $3.96 |
For a batch job that does not need an answer right now, the schedule matters as much as the model.
| Example job | Off-peak | Peak |
|---|---|---|
| Flash: 10 million uncached input and 2 million output tokens | $2.70 | $5.40 |
| Pro: the same job | $10.56 | $21.12 |
Assumption, not a measurement: 10 million uncached input tokens and 2 million output tokens. Cache hits would cost less.
Search and fetch: what 1 November costs
If an agent makes 10,000 search requests a day, billing at $0.75 per 1,000 comes to $7.50 a day, or $225.00 over 30 days. Fetch is billed the same way, separately.
Assumption, not a measurement: 10,000 successful search requests a day. DeepSeek's page says billing is per successful request.
DeepSeek direct, or another host?
DeepSeek-V4.1-Flash-1017 is also served by 29 other hosts on OpenRouter. The chart shows the five cheapest and the two dearest, next to DeepSeek's own prices. Blended price is (3 x input + output) / 4.
Host prices: OpenRouter, read 26 October 2026. DeepSeek's own bars come from its page. * Off-peak is derived by us as half of the peak blended price, which is how DeepSeek's page defines it.
How we know, and what we did not verify
Prices, hours and dates were parsed from the saved rendered text of api-docs.deepseek.com/quick_start/pricing (SHA-256 dfb3d0a8006d276f), and the build stops if the page stops saying what the parser expects. OpenRouter host prices are a separate source read on the same day.
- DeepSeek says it may adjust prices and recommends checking the page. We re-read this page weekly.
- We did not test billing. The search and fetch charge is as the page states it, not as we observed it on an invoice.
- We have not asked DeepSeek to comment before publication. A provider or reader who finds a wrong figure can tell us; we correct it and date the correction.
Part of the series: AI price changes calendar and which plan gets the flagship model.
FAQ
When does DeepSeek start charging for search and fetch?
DeepSeek's pricing page, read on 26 October 2026, says per-request billing starts at 00:00 Beijing time on 1 November 2026, at $0.75 per 1,000 successful requests for Search and $0.75 for Fetch. It says server-side calls made through DeepSeek Harness, Claude Code or Codex are currently free.
What are DeepSeek's peak hours?
DeepSeek says peak hours are 01:00 to 04:00 and 06:00 to 10:00 UTC, Monday to Friday, excluding Chinese public holidays. Every other hour, including weekends and those holidays, is off-peak, and off-peak rates are half the peak rates.
How much cheaper is off-peak?
Exactly half, per DeepSeek. For deepseek-flash, uncached input is $0.15 off-peak against $0.30 at peak, and output is $0.60 against $1.20, per million tokens.
Is deepseek-v4-flash still available?
DeepSeek says that name is still accepted but the model is retired, and requests are served by DeepSeek-V4.1-Flash-1017 and billed at the Flash price. DeepSeek recommends the name deepseek-flash.
Are other hosts cheaper than DeepSeek for the same model?
On OpenRouter, 9 of 29 other hosts of DeepSeek-V4.1-Flash-1017 list a blended price below DeepSeek's own off-peak blended price, with Sail Research lowest at $0.11. Hosts differ in quantisation, speed, uptime and data policy, so a lower price is not the same service.