DeepSeek V4 Pro 0423 API Prices Jump 358%: What It Costs You Now

The Numbers

DeepSeek has raised API pricing on V4 Pro 0423 sharply. Input tokens move from $0.209 to $0.955 per 1M tokens. Output tokens go from $0.418 to $1.91 per 1M tokens. That's a +358% increase across the board — not a rounding error.

Does It Actually Hurt?

It depends entirely on your workload shape.

• Low-volume or prototype users will barely notice. Even at $1.91/1M output tokens, casual API use stays cheap in absolute terms.
• High-throughput production pipelines — summarisation, RAG retrieval, batch classification — will see meaningful cost line movement. If you were running 500M output tokens a month, your bill jumps from roughly $209 to $955 on input alone.
• Output-heavy tasks (long-form generation, code completion with verbose responses) take the harder hit, since output was already 2× input and that ratio holds at the new rates.

What To Do

Before absorbing the increase, it's worth a quick audit:

1. Profile your token split. If your calls are input-heavy and short on output, the damage is more contained.
2. Compare alternatives. Several models in a similar capability tier are available at lower current rates — use the comparison tools on this site to run a side-by-side cost estimate against your actual token volumes.
3. Check caching eligibility. If the provider offers prompt caching, repeated context can blunt input cost spikes.

For current pricing, context window details, and a running price history, see DeepSeek V4 Pro 0423 — live specs & price history.

Price increases of this magnitude are worth a 10-minute audit before your next billing cycle.

Ezra, Scout AI Team