Unverified50% confidenceFactExact time
Qwen3.8 Flash Next原生支持文本、图像和视频多模态处理,原生上下文窗口达26.2万token,借助YaRN技术可扩展至100万token
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/6/2026
First Seen
Valid until: 12/5/2026
Sources
Related Entities
Related Claims
UnverifiedQwen3.8-Flash-Next支持文本与图像混合输入，上下文长度最高可达262K token，提供None、Low、Medium和Extra High四档可调推理强度73% similarUnverifiedQwen-Audio-3.0-TTS提供Flash和Plus两种版本，Flash主打实时交互，Plus主打高质量生成70% similarUnverifiedQwen3.5是一个总参数量3970亿、每次激活170亿参数的MoE视觉语言模型，支持图像、视频等输入并输出文字68% similarUnverifiedQwen3 27B是原生多模态模型，支持视觉输入，弥补了DeepSeek V-Flash无法处理图像的短板68% similarVerifiedQwen3.6-27B支持多模态能力，包括文本推理、图像理解和视频处理67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/866946API
curl https://kongchang.com/api/v1/knowledge/claims/866946MCP
get_claim(id=866946)