deepseek-v4.1-flash API — per-token pricing and code examples

Checking statusPrice version 4 · in effect since Nov 12, 2026

Price per 1M tokens

InputOfficial price $0.15$0.015
Cache readOfficial price $0.003$0.0021
Cache write$0.015
OutputOfficial price $0.60$0.06

Minimum per request$0.00002

Cache rates apply when the provider reports cache tokens.

90% off

About this model

deepseek-v4.1-flash is served through OpenAI Chat Completions, OpenAI Responses, Anthropic Messages. You pay $0.015 per 1M input tokens and $0.06 per 1M output tokens. That is 90% below the official price of $0.15 and $0.60. Cache reads cost $0.0021 per 1M tokens when the provider reports cached tokens. It is included in budget packs from $1.99 up.

Where DeepSeek-V4.1-Flash fits

DeepSeek-V4.1-Flash is, in DeepSeek's words, the smallest model in its new family, first released on September 10, 2026. The facts on this page follow DeepSeek's documentation for the October 17 version, DeepSeek-V4.1-Flash-1017, which DeepSeek's own API serves under the model ID deepseek-flash. DeepSeek's announcement for that version pitches it at building complete projects from scratch and says it is optimized for coding tools such as DeepSeek Harness and Tencent CodeBuddy.

DeepSeek lists a 1M-token context and up to 384K tokens of output, room that helps when an agent writes many files in one response. Unlike DeepSeek-V4-Pro, this model has vision support in DeepSeek's feature table, so you can send images such as screenshots or diagrams along with text. DeepSeek documents text as the output.

Thinking that you can switch off

DeepSeek-V4.1-Flash runs in thinking mode by default, with high as the default effort, and it also has a non-thinking mode. On /v1/chat/completions you switch modes with thinking: {"type": "enabled"} or "disabled" and choose depth with reasoning_effort, which DeepSeek accepts as low, high or max. Other names are mapped: minimal and low become low, medium and xhigh become high, and ultra becomes max.

Sampling works differently here than on most models. DeepSeek says top_p on this model only takes effect in thinking mode, with an effective range of 0.95 to 1.0, and values below 0.95 are treated as 0.95. In non-thinking mode top_p is fixed at 1.0. In thinking mode temperature, presence_penalty and frequency_penalty raise no error but have no effect.

For tool use, DeepSeek asks you to pass back the reasoning_content of all previous turns whenever the request includes tools. According to the same guide, leaving it out in a tool loop gets a 400 error. Without tools, the field is ignored if you send it.

{
  "model": "deepseek-v4.1-flash",
  "thinking": { "type": "disabled" },
  "messages": [{ "role": "user", "content": "Extract the invoice number from this text" }]
}

This is the /v1/chat/completions format. On DeepSeek's own Responses API, DeepSeek documents {"reasoning": {"effort": "none"}} as the way to turn thinking off, and on DeepSeek's own Anthropic-compatible API it documents output_config.effort for depth.

What it costs on Cent AI

You pay per token from a prepaid USD balance, at the four rates shown in this page's price box. With thinking on, DeepSeek-V4.1-Flash spends reasoning tokens before it answers, and those are billed as output. Since thinking is the default, simple extraction or classification calls cost less once you switch thinking off or set reasoning_effort to low.

DeepSeek enables context caching by default for all users, with no code change needed; on DeepSeek's own API, hits are reported as prompt_cache_hit_tokens. DeepSeek calls this best-effort: a hit needs a matching prefix, and cached data is usually cleared within a few hours to a few days. Cent AI charges the cache rates only for the hits the model reports, and shows those cached tokens in the usage of each response in the format of the endpoint you call, for example prompt_tokens_details.cached_tokens on /v1/chat/completions.

Using it from your tools

DeepSeek-V4.1-Flash is reachable on /v1/chat/completions, /v1/responses and /v1/messages. Cent AI does not change your parameters silently, so the effort mapping and the top_p rules above are applied by the model, not by Cent AI. DeepSeek also notes that on DeepSeek's own Anthropic-compatible API, cache_control and budget_tokens are ignored.

Example: the cost of one request

Take a request that sends 2,000 new input tokens, reads 8,000 tokens from cache and gets 800 output tokens back. At the rates in the price box on this page:

  • input: 2,000 × $0.015 / 1,000,000 = $0.00003
  • cache read: 8,000 × $0.0021 / 1,000,000 = $0.0000168
  • output: 800 × $0.06 / 1,000,000 = $0.000048

That request costs $0.0000948, under Rp2 at Cent AI's exchange rate on 2026-11-14 (Rp18,231.96 per USD). A thousand requests like it cost $0.0948. Your real requests will differ; each one shows its own amount in usage.cost and on the Requests page.

Questions

Is there a 5-hour or weekly limit? No. You pay per token from your balance. Each key accepts up to 600 requests per minute, and you can set your own daily or monthly limit on a key.

Is a failed request charged? A request that fails before the model receives it costs $0. Tokens the model already produced are charged.

Does unused balance expire? No, not while your account exists. Every top-up is manual, with no auto-renewal.

Compare with deepseek-v4-pro, glm-5.3-flash, qwen3.8-flash, or see every model on the pricing page.

Specification

Context window
—
Input and output
Text
Endpoints
/v1/chat/completions, /v1/responses, /v1/messages
Smallest pack
$1.99

Latency and error rate per model are on the status page. status.cent-ai.dev

What a request costs

Open the full calculator

Cost per request

90% is an assumption you can change, not a promise. Cache is billed only from cache tokens the model provider reports.

See how many such requests it covers.

Estimate

deepseek-v4.1-flash
Fresh input99,000 tok
Cache read891,000 tok
Output10,000 tok

Per request$0.003956
At official pricesAt official prices $0.0235

Total$0.003956

An estimate at price version 4. Your bill comes from the usage the provider reports for each request.

Save $0.0195

Code

curl https://api.cent-ai.dev/v1/chat/completions \
  -H "Authorization: Bearer $CENT_AI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4.1-flash", "messages": [{"role": "user", "content": "Hello"}]}'

Set CENT_AI_KEY to a key from the dashboard. OpenAI SDKs use https://api.cent-ai.dev/v1; Anthropic SDKs use https://api.cent-ai.dev.

Price history

Version 4, in effect since Nov 12, 2026.

  1. Effective

    per 1M tokens

    Inputold $0.015 new $0.015
    Cache readold $0.0015 new $0.0021
    Cache writeold $0.015 new $0.015
    Outputold $0.06 new $0.06
  2. Effective

    per 1M tokens

    Inputold $0.0075 new $0.015
    Cache readold $0.00075 new $0.0015
    Cache writeold $0.0075 new $0.015
    Outputold $0.03 new $0.06
  3. Effective

    per 1M tokens

    Inputold $0.0075 new $0.0075
    Cache readold $0.00015 new $0.00075
    Cache writeold $0.0075 new $0.0075
    Outputold $0.03 new $0.03

Compare with

All models against official prices

Questions

How much does deepseek-v4.1-flash cost on Cent AI?

$0.015 per 1M input tokens and $0.06 per 1M output tokens. Cache reads cost $0.0021 and cache writes $0.015 per 1M tokens. Each request is rounded up once.

Is there a minimum charge?

A successful request costs at least $0.00002. A request that ends early, because the client left or the output was cut off, pays only for its tokens.

How do I pay?

Top up a prepaid USD balance with QRIS or USDT/USDC, or buy a budget pack once. Nothing renews automatically.

What if the provider does not report usage?

The request is billed from an estimate and marked estimated on the Requests page. An estimate never contains cache tokens.

Which endpoint do I call?

/v1/chat/completions, /v1/responses, /v1/messages. The OpenAI base URL is https://api.cent-ai.dev/v1.

Try deepseek-v4.1-flash with one key

One prepaid balance for every model. Pay with QRIS or USDT/USDC.

Start free