Qwen3.8-Flash-Next发布,6B激活超越Claude Opus 4.6

Chubby♨️ · @kimmonismus · X·2026-08-26 20:44·31分钟前
AI 导读

Qwen3.8-Flash-Next正式发布,这是一款高度稀疏MoE开源多模态模型,总参数125B加51B n-gram嵌入,每token仅激活6B参数,原生256K上下文可经YaRN扩展至1M。

Chubby♨️@kimmonismus
54AI 编辑部评分,满分 100

Qwen3.8-Flash-Next发布,6B激活超越Claude Opus 4.6

2026-08-26 20:44· 31分钟前
AI 导读

Qwen3.8-Flash-Next正式发布,这是一款高度稀疏MoE开源多模态模型,总参数125B加51B n-gram嵌入,每token仅激活6B参数,原生256K上下文可经YaRN扩展至1M。

Qwen 3.8 Flash-Next official released: A 6B-active open model just beat Claude Opus 4.6 Max across 8 of 9 comparable benchmarks!

Qwen3.8-Flash-Next is a highly sparse MoE:

• 125B model parameters • 51B additional n-gram embeddings • Only 6B parameters active per token

It scores: • 62.5 SWE-bench Pro • 81.0 SWE-bench Multilingual • 73.9 CoworkBench • 55.7 JobBench • 73.5 Toolathlon • 81.3 IFBench • 91.7 GPQA Diamond • 91.9 LiveCodeBench

It also outperforms Qwen3.8-27B and DeepSeek-V4-Flash across most of the table.

Super cool release!!

ModelScopeQwen3.8-Flash-Next is here! An open-weight multimodal MoE built on a brand new architecture, with native 256K context extendable to 1M via YaRN. 🤖https://www.m...