【文章标题】：Qwen 3.8 27B Uncensored – Testing the Uncensored Qwen Model 【文章标题】：Qwen 3.8 27B 无审查版 – 测试无审查版 Qwen 模型
【文章正文】： 【文章正文】：
Model guide 模型指南
Qwen 3.8 27B Uncensored overview Qwen 3.8 27B 无审查版概述
This imageat deployment packages Qwen 3.8 27B Uncensored as a hosted chat API: no model download, inference server, or GPU capacity planning is required. Applications send familiar system, user, and assistant messages, while imageat handles execution, credit settlement, and the final OpenAI-style response. 此 imageat 部署将 Qwen 3.8 27B 无审查版封装为托管聊天 API：无需下载模型、配置推理服务器或规划 GPU 容量。应用程序发送熟悉的系统、用户和助手消息，而 imageat 负责执行、积分结算以及最终生成 OpenAI 风格的响应。
The model is useful when a request needs more room than a typical short-context assistant—large specifications, multi-file code excerpts, research notes, or extended conversations. Thinking mode is a per-request choice, so one integration can support both quick chat and more deliberate analysis. 当请求需要比典型短上下文助手更多空间时（如大型规范文档、多文件代码片段、研究笔记或长篇对话），该模型非常有用。思考模式可按请求选择，因此一次集成即可同时支持快速聊天和更深入的分析。
What you can build 您可以构建的应用
-
Add long-context assistance to an imageat workflow without hosting a separate LLM service
-
无需托管独立的 LLM 服务，即可为 imageat 工作流添加长上下文辅助功能
-
Review large product specifications, logs, or code excerpts submitted by your backend
-
审查后端提交的大型产品规范、日志或代码片段
-
Generate structured creative briefs before sending work into image and video pipelines
-
在将任务发送至图像和视频流水线之前，生成结构化的创意简报
-
Build support agents that preserve role-based conversation context across several turns
-
构建支持代理，在多轮对话中保留基于角色的对话上下文
-
Turn research notes into implementation plans, checklists, documentation, or reports
-
将研究笔记转化为实施计划、检查清单、文档或报告
-
Prototype prompts in the playground and reuse the selected parameters through the API
-
在 Playground 中构建提示词原型，并通过 API 复用所选参数
Outputs 输出
The model returns a single assistant response as UTF-8 text in an OpenAI-compatible chat completion response. When thinking mode is enabled, the response may include a dedicated reasoning block before the final answer. 该模型以 UTF-8 文本形式返回单个助手响应，格式兼容 OpenAI 聊天补全响应。启用思考模式时，响应可能会在最终答案前包含一个专门的推理块。
choices[0].message.content choices[0].message.content
Read the generated assistant text from this field. 从此字段读取生成的助手文本。
Model inputs 模型输入
Control the response style, length, repeatability, and conversation identity through the imageat Chat Completions API. 通过 imageat 聊天补全 API 控制响应风格、长度、可重复性和对话身份。
messages messages
Your main instructions, questions, and conversation history as role-based text messages. 您的主要指令、问题和对话历史，以基于角色的文本消息形式呈现。
system message system message
An optional message that sets the assistant’s role, tone, and behavioral rules. 可选消息，用于设置助手的角色、语气和行为规则。
enable_thinking enable_thinking
Enables deeper thinking-style output or disables it for shorter, direct responses. 启用更深度的思考风格输出，或禁用以获得更简短、直接的响应。
user & session_id user 和 session_id
Optional identifiers used to keep separate users and conversation sessions. 可选标识符，用于区分不同的用户和对话会话。
temperature, top_p & top_k temperature、top_p 和 top_k
Randomness and probability-sampling controls for generation. 用于控制生成的随机性和概率采样。
repetition_penalty repetition_penalty
Reduces repeated phrases and generation loops. 减少重复短语和生成循环。
length_penalty length_penalty
Nudges the model toward shorter or longer completions. 引导模型生成更短或更长的补全内容。
min_tokens & max_tokens min_tokens 和 max_tokens
Sets optional minimum and maximum output-token limits. 设置可选的最小和最大输出 token 限制。
stop stop
A string or array of phrases that stops generation at the first match. 字符串或短语数组，在首次匹配时停止生成。
seed seed
An optional integer seed for more repeatable sampled outputs. 可选的整数种子，用于提高采样输出的可重复性。
quantization quantization
Trades a small amount of quality for more efficient inference when enabled. 启用时，以牺牲少量质量为代价换取更高效的推理。
do_sample do_sample
Switches between randomized sampling and more deterministic generation. 在随机采样和更具确定性的生成之间切换。
Recommended settings 推荐设置
Use these presets as a starting point based on the official Qwen 3.8 sampling guidance. 根据官方 Qwen 3.8 采样指南，使用这些预设作为起点。
Reasoning and coding 推理与编码
Enable thinking, then use temperature: 1.0, top_p: 0.95, top_k: 20, and repetition_penalty: 1.0. 启用思考模式，然后使用 temperature: 1.0、top_p: 0.95、top_k: 20 和 repetition_penalty: 1.0。
Direct answers and chat 直接回答与聊天
Disable thinking, then use temperature: 0.7, top_p: 0.80, top_k: 20, and repetition_penalty: 1.0. 禁用思考模式，然后使用 temperature: 0.7、top_p: 0.80、top_k: 20 和 repetition_penalty: 1.0。
Limitations 限制
-
Reduced refusal behavior is not a guarantee that every prompt will be answered or that every answer is appropriate.
-
减少拒绝行为并不能保证每个提示都会得到回答，也不能保证每个回答都恰当。
-
Messages and generated output share the 262,144-token context budget; reserve enough space for the completion.
-
消息和生成的输出共享 262,144 个 token 的上下文预算；请为补全内容预留足够的空间。
-
Large context capacity does not guarantee perfect retrieval—important instructions should be explicit and well structured.
-
大上下文容量不能保证完美的检索——重要指令应明确且结构良好。
-
Thinking mode can materially increase runtime, output length, and the final imageat credit tier.
-
思考模式可能会显著增加运行时间、输出长度以及最终的 imageat 积分消耗层级。
-
Sampling and quantization settings can change factual consistency, formatting, and reproducibility.
-
采样和量化设置可能会改变事实一致性、格式和可重复性。
-
The imageat endpoint documented here accepts text messages; it does not expose the checkpoint’s possible vision inputs.
-
此处记录的 imageat 端点接受文本消息；它不公开模型检查点可能支持的视觉输入。
Safety and compliance 安全与合规
This model may be more willing to answer sensitive requests than strongly aligned chat models. You must still follow imageat platform rules and all applicable laws. Do not use it for instructions that enable wrongdoing, violence, self-harm, or illegal access. 与经过严格对齐的聊天模型相比，该模型可能更愿意回答敏感请求。您仍必须遵守 imageat 平台规则和所有适用法律。请勿将其用于促成不当行为、暴力、自我伤害或非法访问的指令。
For user-facing products, add your own moderation, logging, rate limits, abuse prevention, and human review for high-risk use cases. 对于面向用户的产品，请添加您自己的内容审核、日志记录、速率限制、防滥用机制，并对高风险用例进行人工审查。