模型介绍
Qwen-Image's text encoder, Qwen3-VL-8B-Instruct, with its refusal direction projected out: a drop-in --llm for stable-diffusion.cpp next to our Turbo denoiser GGUFs . Below, exactly which part is uncensored and what that does to the pictures. The measurements are public. As a chat model, the encoder's refusals fall from 88.9% to 1.2% on held-out English prompts. MMLU is unchanged: 77.35% before and after. On images it changes nothing we could measure. The stock Qwen-Image pipeline has no filter and already draws all 9 sensitive categories we tested, and with this encoder not one of 45 prompts changed outcome. See where exactly this is uncensored . Generate images locally in Atomic Chat . It runs stable-diffusion.cpp, the engine these files were built and checked with. Qwen-Image 2.1 makes a picture with three parts. Only one of them is changed here.
主要特点
特点摘要待依据来源补齐，不凭模型名称推断能力。
局限与注意事项
暂无足够来源支持具体缺点结论。请重点验证实际任务表现、时延、数据政策、许可与版本生命周期。
来源报价与费用分析
同一模型在不同渠道、地区和配置下可能采用不同价格。下表独立保留各渠道报价，不将渠道价格视为厂商官方统一价格。
| 来源 / 渠道 | 输入 | 输出 | 计费单位 | 采集时间 |
|---|---|---|---|---|
| 暂未取得可核验报价，请查阅来源渠道。 | ||||
按月估算基础 Token 费用
预估不包含缓存优惠、阶梯定价、多模态附加费用、网络检索、税费及推理额外输出。
公开评测与运行配置
暂无已采集的可核验公开评分。模型未上榜不代表能力较低。
查看完整已收录榜单 →常见问题
开放权重是否代表可以免费商用？
不一定。应核对具体版本许可证、商用门槛和使用政策。自部署还会产生算力和维护成本。
这些价格是实时结算价格吗？
不是实时结算价格，而是带采集时间的来源报价快照。请到对应渠道核对当前版本、区域、折扣和计费方式。
资料来源与可追溯记录
许可证：apache-2.0。模型介绍与第三方评测均以原来源的具体版本和方法为准。