Models · DeepSeek

DeepSeek V4 Flash pricing: the cheapest row, with an asterisk

An order of magnitude below everything else, with a 98% cache discount — and a data-residency question you have to answer before the price matters.

All numbers computed from pricing data verified 2026-07-11 — this page updates with every pricing refresh.

The rate card

ModelInput /1MOutput /1MCached readCache writeBatch inputContext
DeepSeek V4 Flash$0.14$0.28$0.0028$0.071000K

What it costs on real workloads

Realistic monthly cost at 1,000 requests/day with the calculator’s defaults — 8% retries, 40% cache hit rate, 15% infra overhead, per-model tokenizer calibration applied. The same three workloads used across every comparison on this site, so the numbers line up.

WorkloadDeepSeek V4 Flash
Simple chatbot$7/mo
Chatbot with history$16/mo
Multi-step agent$88/mo

The cheapest row in the catalog, by a distance

On the history-carrying chatbot above, DeepSeek V4 Flash runs $16/mo. GPT-5.4 costs roughly 31× that for the same workload; even Gemini 2.5 Flash, the next cheapest model here, is about 4.5× more. At $0.14/1M input and $0.28/1M output, cost effectively stops being the constraint on your architecture.

A 98% cached-read discount — the steepest here

Cached input bills at $0.0028/1M, roughly 98% below standard input. Every other model in the catalog discounts cached reads by about 90%; this one is closer to 98%. On a workload with a large stable prefix — a long system prompt, a reused document, a growing transcript — the cached portion of your bill rounds to nothing.

The 1000K context window compounds that: long prompts are both cheap to send and nearly free to re-send.

The caveat you have to price in

DeepSeek is a China-based provider, and for many teams that is a decision made outside the cost model entirely — by a procurement policy, a customer contract, a regulator, or a data residency commitment. If you handle personal data, regulated data, or anything a customer agreement covers, review the data policy before the price, because no per-token discount survives a compliance objection.

Where it fits cleanly: internal tooling, non-sensitive bulk processing, prototyping, and offline pipelines over public or synthetic data. Where it usually doesn’t: anything customer-facing under a data-processing agreement that doesn’t list it.

Endpoint naming

The legacy endpoint names deepseek-chat and deepseek-reasoner were retired on 2026-07-24; both had already been routing to V4 Flash beforehand. Use deepseek-v4-flash. If you are reading an older integration guide, that is the substitution to make.

Quality is the open question, not price. Our recommendation engine keeps DeepSeek out of chatbot recommendations unless it is more than 50% cheaper than the next non-DeepSeek option — a deliberate data-policy hedge — and excludes budget-tier models from coding and routing archetypes entirely. Validate on your own task before scaling.

Model DeepSeek V4 Flash against the rest of the catalog on your own workload.

Open the calculator →

Related