Claude Fable 5.2 now costs less. Improvements to our serving architecture have lowered the cost of using the model, and the new rates are live.See new pricing
DeepSeek

DeepSeek V4 Pro

Official

deepseek/deepseek-v4-pro

DeepSeek V4 Pro — V4 flagship reasoning model, 1M context, ultra-low cost.

Live model playground

Try DeepSeek V4 Pro here

Tune the request, see the current price ceiling, and inspect the real response without leaving this model page.

streamchat
1Configure input
200 / 20,000
Maximum estimated cost—
2Live resultExample
DeepSeek V4 Pro

I'm sorry for the surprise charge. I'll review your usage details to find out what caused the increase. Could you please provide the email address on your account?

Example output · generated Oct 24, 2026 · took 5.8s · cost $0.001043
Context window
1M
Max output
384K
Input modalities
Text
Output modalities
Text
Capabilities
Reasoning · Tool calling · Coding · Prompt cache
Upstream release
2026-08-13
RouteMux name
deepseek/deepseek-v4-pro

What is DeepSeek V4 Pro

DeepSeek V4 Pro is the larger model in DeepSeek's V4 line. The V4 preview announcement puts it at 1.6T total parameters with 49B active, using token-wise compression plus DSA (DeepSeek Sparse Attention). It shipped as an open-weight preview on April 24, 2026 and reached general availability on August 13; the current build is DeepSeek-V4-Pro-0813, and the API name has stayed deepseek-v4-pro throughout. DeepSeek describes the GA release as a major agent upgrade with native OpenAI Responses API support.

It has a 1M-token context window, up to 384K output tokens, and text-only input: DeepSeek does not offer vision on V4 Pro. Thinking and non-thinking modes share one id, with thinking on by default. Effort has three real levels, low, high and max, which DeepSeek suggests for simple tasks, everyday agent workflows and complex tasks respectively; medium is treated as high. In thinking mode, temperature, presence_penalty and frequency_penalty are ignored without an error, and top_p is ignored too; your top_p only applies in non-thinking mode. It is callable over OpenAI Chat Completions, OpenAI Responses and Anthropic Messages as deepseek/deepseek-v4-pro.

When V4 Pro is still the right pick: DeepSeek upgrades deepseek-flash in place (it moved to V4.1-Flash-1017 on October 17), while deepseek-v4-pro has stayed on the 0813 build since August 13, which suits workflows that want the model version to hold still. It also fits text-only work that relies on top_p in non-thinking mode. When V4.1 Flash launched, DeepSeek planned to route V4 Pro requests to it, then updated its changelog to keep serving the V4 Pro API after September 14 with billing unchanged.

On RouteMux, V4 Pro costs $1.32 per 1M input tokens, $3.96 per 1M output tokens and $0.044 per 1M cached input tokens.

DeepSeek V4 Pro API pricing

Prices verified 2026-06-08

InputOfficial $1.32 / M tokens
$1.32/ M tokens
OutputOfficial $3.96 / M tokens
$3.96/ M tokens

Cache read $0.044/M · Cache write —/M

Estimate cost

Estimated cost$0.01716

Key capabilities

1.6T total, 49B active
The size DeepSeek gave in the V4 preview announcement, with token-wise compression and DSA sparse attention.
Three effort levels
low for simple tasks, high for everyday agent workflows, max for complex tasks. medium maps to high.
1M context, 384K output
A 1M-token context window and up to 384K output tokens per request.
Pinned at 0813
deepseek-v4-pro has not changed builds since the August 13 GA, unlike deepseek-flash, which is upgraded in place.
Text only
DeepSeek does not offer vision on V4 Pro. For image input, use V4.1 Flash.

Limited time10% extra Credits on your first top-up — up to 200 bonus Credits

Applies to your first top-up within 72 hours of signing up. Minimum $20, once per account.

Sign up and claimNo subscription. Pay for what you use.

Interface and protocols

Use any of these IDs to call this model via the API.

RouteMux namedeepseek/deepseek-v4-pro

Replace the ROUTEMUX_KEY placeholder with your API key. Create one →

from openai import OpenAI

client = OpenAI(
    base_url="https://api.routemux.com/v1",
    api_key="<ROUTEMUX_KEY>",
)

completion = client.chat.completions.create(
    model="deepseek/deepseek-v4-pro",
    messages=[{"role": "user", "content": "What is the meaning of life?"}],
)
print(completion.choices[0].message.content)
base_url: https://api.routemux.com/v1

Performance (last 30 days)

DeepSeek V4 Pro FAQ

›What is DeepSeek V4 Pro?

DeepSeek V4 Pro — V4 flagship reasoning model, 1M context, ultra-low cost.

›How large is the context window of DeepSeek V4 Pro?

DeepSeek V4 Pro supports a context window of up to 1M tokens, with up to 384K output tokens per request.

›How much does the DeepSeek V4 Pro API cost?

Through RouteMux, DeepSeek V4 Pro costs $1.32 for input and $3.96 for output (USD per 1M tokens), below the official list price.

›How do I call DeepSeek V4 Pro via API?

DeepSeek V4 Pro is OpenAI-compatible: point your base URL at RouteMux and set the model field to deepseek/deepseek-v4-pro — no code changes needed.

›Which input and output modalities does DeepSeek V4 Pro support?

DeepSeek V4 Pro accepts text as input and produces text as output.

›When was DeepSeek V4 Pro released?

DeepSeek V4 Pro was released on August 13, 2026.

Sources