DeepSeek model API pricing

DeepSeek deepseek-v4-pro API pricing

DeepSeek deepseek-v4-pro API pricing is $0.435 per 1M input tokens and $0.870 per 1M output tokens in the ByteCosts provider pricing index. For an equal 1M input plus 1M output token plan, the listed token cost is $1.30. Context window: 1M. Prompt cache prices: read $0.0036, write -, 1h write - per 1M tokens. The confidence grade is A+: Read from the provider's official pricing page. The source was last checked 2026-07-19. At a reference workload of 1M input tokens and 250K output tokens per month, the listed token cost is $0.652 / month. This is a planning estimate; real invoices can differ because taxes, discounts, commitments, gateway fees, and non-token charges are excluded.

Open the pricing index - DeepSeek / deepseek-v4-pro →

Pricing snapshot

DeepSeek deepseek-v4-pro pricing fields
FieldValue
ProviderDeepSeek
Modeldeepseek-v4-pro
Input / 1M tokens$0.435
Output / 1M tokens$0.870
Cache read / 1M tokens$0.0036
Cache write / 1M tokens-
Cache write 1h / 1M tokens-
Context window1M
Reference monthly workload$0.652 / month

The reference workload is 1M input tokens plus 250K output tokens per month. It excludes taxes, discounts, gateway fees, and non-token charges.

Reference workload examples

Each example is pure arithmetic from this committed record's input and output token prices. The assumptions are shown inline.

DeepSeek deepseek-v4-pro monthly workload examples
ExampleAssumed monthly tokensEstimated token cost
Light reference1M input + 250K output tokens/month$0.652 / month
Production reference10M input + 2.5M output tokens/month$6.52 / month
Heavy reference100M input + 25M output tokens/month$65.25 / month

Examples exclude taxes, discounts, commitments, gateway fees, cache effects, and non-token charges unless those charges are part of the committed token price record.

deepseek-v4-pro pricing across providers

deepseek-v4-pro is listed by 4 providers in the ByteCosts index. Per-million-token list prices, lowest input first.

deepseek-v4-pro input and output prices by provider
ProviderInput / 1MOutput / 1M
DeepSeek (this page)$0.435$0.870
Deep Infra$1.30$2.60
NovitaAI$1.60$3.20
Vivgrid$1.74$3.48

When this model is cost-effective

deepseek-v4-pro ranks 2 of 2 DeepSeek records by combined list price ($1.30 for 1M input plus 1M output tokens). That puts it in the middle third of this provider's committed records.

Cached input reads are listed at $0.0036 per 1M tokens, 0.8% of the fresh input rate, so repeated cached prompts have a lower committed input-token cost than fresh prompts.

The committed context window is 1M; use that as the only long-context signal on this page, not as a quality or latency claim.

Cheaper alternatives

These records have a lower committed input+output list price than this page's record. Same-provider records and same-model cross-provider records are preferred when available.

DeepSeek deepseek-v4-pro lower-priced alternatives
RecordInput / 1MOutput / 1MCombined / 1M + 1MDetail page
DeepSeek deepseek-v4-flash$0.140$0.280$0.420/tools/ai-provider-pricing/deepseek--deepseek-v4-flash
Z.AI GLM-4.7-Flash$0$0$0/tools/ai-provider-pricing/zai--glm-4.7-flash
Z.AI GLM-4.5-Flash$0$0$0/tools/ai-provider-pricing/zai--glm-4.5-flash

Evidence and freshness

ByteCosts grades this row A+. Read from the provider's official pricing page. The record was first tracked on 2026-05-30 and last checked on 2026-07-19.

The source URL is kept with the record so readers can verify the provider page before making a purchasing or architecture decision. Prices change often, so quote the last-checked date with any copied number.

Watch-outs

  • No hidden-cost flags are attached to this source-backed record.

Price history

DeepSeek deepseek-v4-pro tracked changes
WhenFieldWasNowChange
2026-07-09input$0.0036$0.43511900% up
2026-07-09cachedInput$0.015$0.003675% down
2026-06-01cacheReadnew$0.0036new

Frequently asked questions

Is deepseek-v4-pro free to use?

This ByteCosts row is not marked free. Use the listed token rates and verify the provider source for discounts, commitments, and account-specific terms.

Does this page include all possible charges?

No. The page focuses on source-backed token pricing and known cache/context fields. It does not include taxes, enterprise commitments, gateway markups, storage, retrieval, fine-tuning, or platform fees unless those charges are present in the source-backed row.

DeepSeek deepseek-v4-pro API pricing. ByteCosts. Updated 2026-07-19. https://bytecosts.com/tools/ai-provider-pricing/deepseek--deepseek-v4-pro/

Sources

Machine-readable