DeepSeek V4.1 Flash API Prices Jump 1400%: What Developers Need to Know

The Numbers

DeepSeek has repriced V4.1 Flash significantly. Input tokens move from $0.02 to $0.30 per 1M tokens — a +1400% increase. Output tokens double, going from $0.60 to $1.20 per 1M tokens.

The model was previously one of the cheapest capable options on the market. That positioning no longer holds.

Does This Actually Hurt You?

It depends heavily on your workload shape:

• Low-volume or internal tooling: At $0.30 input, you'd pay $0.30 for every million input tokens consumed. For most hobby projects or low-traffic apps, the absolute dollar impact stays small.
• High-volume, input-heavy pipelines (document processing, RAG, batch classification): This is where the 1400% jump bites hard. A pipeline that cost $20 per billion input tokens now costs $300.
• Output-heavy workloads: The output price doubling from $0.60 to $1.20 is painful but proportionally less dramatic than the input shock.

What To Do Now

Before accepting the new rate, it's worth auditing your token split. If your workload is input-heavy, the repricing hits disproportionately. If output-heavy, the damage is more contained.

Alternatives worth checking: other sub-$0.50 input models indexed on our site may now undercut V4.1 Flash on price while remaining competitive on capability. Run a quick cost projection against your actual token ratios before deciding whether to stay or switch.

Full current specs and a live price history are on the DeepSeek V4.1 Flash — live specs & price history page.

Bottom Line

This is a substantial repricing, not a rounding error. High-volume users should model the cost delta immediately. Casual users can probably absorb it without noticing.

Ezra, Scout AI Team