DeepSeek V4.1 Flash API: AI大規模言語モデル

Use DeepSeek V4.1 Flash through MuAPI for multimodal reasoning, image understanding, and adjustable thinking effort. DeepSeek V4.1 Flash is a multimodal reasoning model with a 1M-token context window, image understanding, and adjustable reasoning effort.

Interactive model controls

DeepSeek V4.1 Flash is a multimodal reasoning model with a 1M-token context window, image understanding, and adjustable reasoning effort.

📝

概要

このモデルについて

DeepSeek V4.1 Flash is a fast multimodal reasoning model with a 1M-token context window, native image understanding, and configurable reasoning effort.

1Multimodal image understanding and visual question answering

2Fast reasoning, coding assistance, and long-context analysis

3Agent workflows that benefit from adjustable thinking effort

💰

料金と価値

コスト分析

プロバイダー	費用	備考
muapiapp	Official DeepSeek token rates; peak/off-peak and cache-sensitive	Off-peak: $0.15/M uncached input, $0.003/M cached input, $0.60/M output. Peak rates are 2×. A $0.0001 minimum reservation is reconciled after usage.

muapiappOfficial DeepSeek token rates; peak/off-peak and cache-sensitive

Off-peak: $0.15/M uncached input, $0.003/M cached input, $0.60/M output. Peak rates are 2×. A $0.0001 minimum reservation is reconciled after usage.

** 競合サービスの料金は類似のモデル構成および利用ティアに基づいて算出された推定値です。

⚙️

技術詳細

設定スキーマ

パラメータ	型	説明	デフォルト
プロンプト	string	生成プロセスを適切に制御するためのパラメータ設定（The user message or instruction for the model.）。	`Explain how a mixture-of-experts model works.`
画像 URL	string	Optional public image URL for multimodal analysis.	`undefined`
System プロンプト	string	Optional system-level instruction.	`You are a careful software engineering assistant.`
Reasoning Effort	Enum（7個の選択肢）	Thinking effort. Omit to use the model default; none and minimal disable thinking.	`-`

プロンプトstring

生成プロセスを適切に制御するためのパラメータ設定（The user message or instruction for the model.）。

デフォルト値Explain how a mixture-of-experts model works.

画像 URLstring

Optional public image URL for multimodal analysis.

デフォルト値undefined

System プロンプトstring

Optional system-level instruction.

デフォルト値You are a careful software engineering assistant.

Reasoning EffortEnum（7個の選択肢）

Thinking effort. Omit to use the model default; none and minimal disable thinking.

デフォルト値-

📖

実装ガイド

開発者ドキュメント

How to Use DeepSeek V4.1 Flash

Send a Prompt: Submit a POST request to /api/v1/deepseek-v4-1-flash or stream via /api/v1/deepseek-v4-1-flash/stream.
Add an Image: Optionally provide a public image_url for visual understanding.
Tune Reasoning: Set reasoning_effort to none, minimal, low, medium, high, xhigh, or max; omit it to use the model default.
Read the Result: Receive generated text or a live SSE stream.

❓

よくある質問

FAQ

Does DeepSeek V4.1 Flash support image input?

Yes. You can provide an optional public image URL alongside your prompt.

Can I control thinking effort?

Yes. Set `reasoning_effort` to one of the supported levels. `none` and `minimal` disable thinking.

Does DeepSeek V4.1 Flash support streaming?

Yes. Live streaming is available at `/api/v1/deepseek-v4-1-flash/stream`.