Models·DeepSeek

DeepSeek V4 Pro 0813

deepseek/deepseek-v4-pro-0813

DeepSeek's open-weight reasoning model with tool support and a 1,048,576-token context window.

$0.66Input, per million tokens
$1.98Output, per million tokens
1.05MContext window
393KLongest answer

1.05M tokens of context — about 786,432 words, on a scale that ends at 2M.

What it can do

  • Reasoning
  • Tool use
  • Open weights
  • Takes text
  • Gives text

Released 12 Aug 2026 · checked 3 Oct 2026.

About it

DeepSeek V4 Pro 0813 is a text-only DeepSeek model released on August 12, 2026. It supports reasoning and tools, has open weights, and can handle up to 1,048,576 input tokens at once. Usage costs $0.66 per million input tokens and $1.98 per million output tokens, with a maximum output of 393,216 tokens.

Review

In a line

DeepSeek V4 Pro 0813 is DeepSeek's GA release of DeepSeek V4 Pro. It is a large-scale mixture-of-experts model with open weights, released August 12, 2026.

It takes text in and returns text.

What it's good at

The model supports reasoning, so it suits tasks that need step-by-step thought. It also supports tools, allowing connected tool calls where the surrounding app provides them.

Its open weights make it relevant to people comparing open-weight models.

Price in plain terms

Input costs $0.66 per million tokens, and output costs $1.98 per million tokens. Tokens are the small pieces of text counted for billing.

Using the rough figures supplied, reading a 500-word page costs about 0.043 cents. A chat turn with about 1,000 input tokens and 500 output tokens costs about 0.17 cents; 1,000 such turns cost about $1.65.

Context and limits

It can take up to 1,048,576 tokens in one request, giving it room for very large text inputs. A single answer can contain up to 393,216 tokens.

The model is text-only: its listed inputs and outputs are text, not images.

Ways to use it

Use it for text conversations and other text tasks where reasoning or tool support matters. Its model identifier is deepseek/deepseek-v4-pro-0813.

Because the weights are open, it also belongs in comparisons of DeepSeek models and reasoning models.

Verdict

Choose this model if you need a reasoning-focused text model, very long context, tool support, and access to open weights. Its low rough chat cost may suit high-volume text use.

Look elsewhere if you need image handling or a model with non-text inputs or outputs. Buyers should also account for separate input and output token charges rather than treating the model as a flat-price service.

Verdict

DeepSeek V4 Pro 0813 fits readers who want open weights, reasoning, tool support, and an unusually large text context. Its token rates are low in the supplied examples, including about $1.65 for 1,000 typical chat turns. The trade-off is a text-only design: image work is outside its listed inputs and outputs. It is a practical choice for long documents and text conversations, provided your software can use this model and you are comfortable with usage-based billing.

  • Open weights are available.
  • Handles up to 1,048,576 input tokens.
  • Supports reasoning and tools.
  • About $1.65 for 1,000 rough chat turns.
  • Text inputs and outputs only.
  • Output is billed separately at $1.98 per million tokens.
  • Maximum output is 393,216 tokens.
  • No image capability is listed.
  • Pick it ifLong text analysis and conversations; Reasoning tasks with tool support; Users comparing open-weight models
  • Skip it ifYou need image input or image output.; You prefer a flat subscription instead of token-based charges.

Pricing

Usage is billed separately for text read and text generated: $0.66 per million input tokens and $1.98 per million output tokens. The supplied examples put a typical chat turn at about 0.17 cents and 1,000 such turns at about $1.65.

Questions

What is DeepSeek V4 Pro 0813?

It is DeepSeek's GA release of DeepSeek V4 Pro, a large-scale mixture-of-experts text model released August 12, 2026.

Is DeepSeek V4 Pro 0813 open weight?

Yes. Its weights are listed as open.

How much does it cost?

Input costs $0.66 per million tokens and output costs $1.98 per million tokens.

How much text can it read?

Its context window is 1,048,576 tokens.

How long can one answer be?

The maximum output is 393,216 tokens.

Does it support reasoning and tools?

Yes. The facts list both reasoning and tool support.

Can it handle images?

No image handling is listed. Its inputs and outputs are text.

On these boards

More from DeepSeek