Qwen3 safety-moderation series
According to the model card, Qwen3Guard is built on Qwen3 and trained on 1.19 million safety-labeled prompts and responses across 0.6B, 4B, and 8B sizes.
Open Source Model Profile · Qwen
Qwen3Guard-Gen-8B
Qwen3Guard-Gen-8B is an 8.19B-parameter Qwen3 safety-moderation model from Qwen with three-tier severity classification.
- Publisher
- Qwen
- Task
- text-generation
- Model type
- qwen3
- License
- apache-2.0
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
Qwen3Guard-Gen-8B is published by Qwen as a qwen3 text-generation safety model. The captured configuration identifies Qwen3ForCausalLM and Safetensors metadata reports 8,190,735,360 parameters. According to the model card, it is the generative Qwen3Guard variant for instruction-style safety classification with safe, controversial, and unsafe severity levels, and card data records apache-2.0.
Recorded capabilities
Three-tier severity classification
According to the model card, outputs are categorized as safe, controversial, or unsafe with category labels such as Violent, PII, and Jailbreak.
119-language support
According to the model card, the Gen variant supports 119 languages and dialects for cross-lingual moderation.
Documented moderation workflows
According to the model card, Transformers and OpenAI-compatible server examples cover both prompt and response moderation with label extraction.
Use cases in the source record
- Prompt and response safety moderation with safe, controversial, and unsafe severity labels plus category extraction.
- Multilingual moderation experiments across the publisher-described 119 supported languages and dialects.
Limitations and unknowns
- No evaluation results were extracted from this record as structured benchmarks.
- No context-window value was extracted from this record.
- Provider state is historical snapshot data and should be refreshed before being presented as current.
- Benchmark superiority language comes from the publisher model card and was not independently verified by Ethen.
Source and provenance
Source: Qwen/Qwen3Guard-Gen-8B
Captured: Unknown. Processed: 2026-09-07T19:34:35.958847+00:00.
Qwen3Guard-Gen-8B Qwen3Guard is a series of safety moderation models built upon Qwen3 and trained on a dataset of 1.19 million prompts and responses labeled for safety. The series includes models of three sizes (0.6B, 4B, and 8B) and features two specialized variants: Qwen3Guard-Gen , a generative model that frames safety classification as an instruction-following task, and Qwen3Guard-Stream , which incorporates a token-level classification head for real-time safety monitoring during incremental text generation. This repository hosts Qwen3Guard-Gen , which offers the following key advantages: Three-Tiered Severity Classification: En…
F001F002F003F004F005F006F007F009F010F011F012F013F014F015F019F020