DeepSeek Warned Prices Are Going Up. It Won't Say How Much or When.

Published: August 12, 2026 Impact: High. Affects every developer and business currently budgeting against DeepSeek's API rates.

What DeepSeek Posted

On August 6, 2026, DeepSeek added a footnote to its developer pricing page. It reads, verbatim:

"We plan to raise the overall pricing for DeepSeek API services in the near future, with a significant increase expected. Please plan your usage accordingly. The specific pricing plan will be subject to official notice."

Bloomberg and the South China Morning Post both covered the notice the same day. Neither the notice nor the coverage contains a number or an effective date.

A price increase announced without a figure or a date is not a disclosure. It is a notice that a disclosure is coming, telling customers to plan around information they have not been given.

I Checked Again Today

I rechecked api-docs.deepseek.com/quick_start/pricing directly today, August 12, twice. Six days after the notice went up, the page still carries no number and no effective date. Current rates stand as the pre-hike baseline: DeepSeek V4-Flash at $0.14/$0.28 per million input/output tokens, V4-Pro at $0.435/$0.87.

Credit where it's due: DeepSeek didn't have to say anything at all. A silent repricing is the industry norm. Pre-announcing, even vaguely, gives customers more warning than most vendors bother to give. But "plan your usage accordingly" is not something a developer can act on. It names an outcome without giving anyone the inputs to plan against it.

An Eleven-Day-Old Conclusion, Now Time-Limited

On August 1, I published BNA #18, concluding that DeepSeek is still cheaper than OpenAI's newly cut Luna API rate on both input and output. That comparison was accurate the day it published. It now carries an expiration date nobody, including DeepSeek, has named. I'm not retracting the finding. I'm flagging that the ground under it can move at any time, without warning of size or timing.

Naming My Own Exposure

OneHuman's recommendation engine currently ranks DeepSeek first in low-budget brackets, a deliberate, disclosed price-fit weighting, not a bug. A pending, undated increase on the exact tool that engine recommends on price is a cheap audit opportunity I'd rather name now than after the fact.

The Number That Is Verified: 82 Days

DeepSeek's current API rate has held flat for 82 days, since a permanent price cut took effect on May 22, 2026. OneHuman's own June 15 pricing sweep independently confirmed the same figures. That is the real, checkable window: 82 days of stable pricing, nothing longer claimed. This field only carries a clean, verified reading for this specific stretch, so this is the exact stretch I'm reporting, not an estimate stretched further back.

What to Do Next

If you're budgeting against DeepSeek's API today: the $0.14/$0.28 and $0.435/$0.87 rates are real and current, but treat them as provisional, not locked.

If you're deciding whether to migrate off DeepSeek: you cannot make that call rationally yet. Neither can DeepSeek's own customers. Wait for the actual number before treating this as a reason to move.

If you're comparing DeepSeek to OpenAI's Luna rate: the August 1 comparison holds today. Recheck it once DeepSeek publishes real figures.

Sources

• DeepSeek API pricing page, footnote 2, checked August 6 and independently rechecked August 12, 2026
• Bloomberg and South China Morning Post, DeepSeek pricing notice coverage, August 6, 2026
• OneHuman BNA #18, "OpenAI Cut Luna's API Price 80%. DeepSeek Is Still Cheaper." (August 1, 2026)
• OneHuman pricing sweep, June 15, 2026 (internal): V4-Pro $0.435/$0.87, V4-Flash $0.14/$0.28, confirming the May 22 permanent cut

Verified by OneHuman · August 12, 2026

Independent. AI-assisted. Human-verified. No ads. No affiliates. No investors.