muapi Logomuapi Logo
muapi Logomuapi Logo

DeepSeek V4.1 Flash Abliterated: Большие языковые модели искусственного интеллекта

Call DeepSeek V4.1 Flash Abliterated through one API. 1M-token context, thinking mode, streaming and pay-per-token pricing. Try free in the Muapi playground.

Интерактивное управление моделью

DeepSeek V4.1 Flash Abliterated is a fast DeepSeek V4.1 Flash reasoning model with a 1M-token context window and tool-calling support, modified with abliteration to reduce refusal behavior; an optional thinking mode enables extended reasoning.

📝

Обзор

Об этой модели

DeepSeek V4.1 Flash Abliterated is a fast DeepSeek V4.1 Flash reasoning model with a 1M-token context window and tool-calling support, modified with abliteration to reduce refusal behavior; an optional thinking mode enables extended reasoning. Muapi exposes it through a single REST endpoint with async polling and Server-Sent Events streaming, so you can drop it into chat apps, agents and workflows without hosting GPUs. It is best suited to long-context reasoning, coding and tool-using agents. Pricing is pay-per-token at $0.66/M input tokens and $2.64/M output tokens, with no subscription. See also GLM 5.3 Flash Abliterated, DeepSeek R1 Llama 70B Abliterated, DeepSeek R1 Qwen 32B Abliterated. Browse the full lineup on the abliterated LLM API hub; related families: GLM Abliterated API and DeepSeek R1 Abliterated API. 'Abliterated' and 'derestricted' describe modifications that reduce the model's refusal behavior. They do not guarantee unrestricted output, and you remain responsible for complying with applicable laws and the Muapi terms of use.

1Chat and assistants: Build conversational apps on DeepSeek V4.1 Flash Abliterated for long-context reasoning, coding and tool-using agents.
2Low-refusal experimentation: Compare refusal behavior against the base model in your own evaluations, within Muapi's usage policy.
3Agents and workflows: Call the model from Muapi workflow nodes and the [abliterated LLM API](/abliterated-llm-api) hub examples.
💰

Цены и стоимость

Анализ затрат

muapiapp$0.66/M input tokens and $2.64/M output tokens

Оплата за токены по фактическому использованию; при запуске резервируется небольшая минимальная сумма.

Fal.aiНедоступно

DeepSeek V4.1 Flash Abliterated is not offered as a hosted endpoint.

ReplicateНедоступно

DeepSeek V4.1 Flash Abliterated is not offered as a hosted endpoint.

** Цены конкурентов рассчитаны на основе аналогичных архитектур моделей и уровней использования.

⚙️

Технические детали

Схема конфигурации

подсказатьstring

.

Значение по умолчаниюSummarize the trade-offs between a monolith and microservices in five bullet points.
Системная подсказкаstring

, .

Значение по умолчаниюYou are a concise, direct assistant.
Максимальное количество токеновinteger

Optional maximum number of tokens to generate.

Значение по умолчанию8192
мышлениеboolean

Enable extended thinking mode for harder reasoning tasks.

Значение по умолчаниюfalse
📖

Руководство по внедрению

Документация разработчика

How to Use DeepSeek V4.1 Flash Abliterated API

  1. Get an API key: Create a key in your Muapi dashboard.

  2. Send a prompt: POST prompt (plus optional system_prompt and max_tokens, up to 32768) to /api/v1/deepseek-v4-1-flash-abliterated. Set thinking to true for extended reasoning.

curl -X POST https://api.muapi.ai/api/v1/deepseek-v4-1-flash-abliterated \
  -H "x-api-key: $MUAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Explain how attention works in transformers.", "thinking": true}'
  1. Poll or stream: Poll GET /api/v1/predictions/{request_id}/result until status is completed, or stream tokens live from /api/v1/deepseek-v4-1-flash-abliterated/stream.
curl -X POST https://api.muapi.ai/api/v1/deepseek-v4-1-flash-abliterated/stream \
  -H "x-api-key: $MUAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Explain how attention works in transformers."}'
  1. Read the result: The generated text is returned in outputs. Usage is billed on actual token counts.
❓

Общие вопросы

Часто задаваемые

What is DeepSeek V4.1 Flash Abliterated?

DeepSeek V4.1 Flash Abliterated is a fast DeepSeek V4.1 Flash reasoning model with a 1M-token context window and tool-calling support, modified with abliteration to reduce refusal behavior; an optional thinking mode enables extended reasoning.

What does abliterated mean?

Abliterated or derestricted models are modified to reduce refusal behavior. This lowers, but does not eliminate, refusals and does not guarantee unrestricted output.

How is DeepSeek V4.1 Flash Abliterated priced?

Billing is per token: $0.66/M input tokens and $2.64/M output tokens. A small minimum is reserved when the request starts and reconciled to actual usage.

Does DeepSeek V4.1 Flash Abliterated support streaming?

Yes. Stream tokens live from `/api/v1/deepseek-v4-1-flash-abliterated/stream` using Server-Sent Events, or poll the async endpoint.

How do I turn on thinking mode?

Send `"thinking": true` in the request body to use the reasoning variant. Leave it out for the standard, faster variant.

What is the context window?

The context window is about 1M tokens. Prompt and output together must fit inside it; `max_tokens` is capped at 32768. The endpoint accepts text input only.

Is output guaranteed to be unrestricted?

No. Abliterated models are less likely to refuse, but behavior varies by prompt and quality can differ from the base model. Evaluate it for your use case and follow applicable laws.