阿里巴巴的Qwen团队发布了Qwen3.8-Flash，这是一个开放权重的多模态模型，作为Qwen4架构的早期预览。该新模型在成本效益和性能方面都有显著提升，在各种基准测试中均优于其前身Qwen3.7-Plus，尤其是在编码和办公任务方面。关键架构升级包括混合注意力机制、门控残差网络和N-gram嵌入系统，这些都为其增强的功能和降低的计算成本做出了贡献。 AI
影响 在多个基准测试中设定了新的SOTA（State-of-the-Art），重点关注成本效益，可能影响未来的模型开发和部署策略。
排序理由 前沿实验室模型发布，附带系统卡和基准测试结果。
AI 生成摘要 · Google Gemini · 来自 5 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
With only 6B active parameters, Qwen3.8-Flash-Next-Base tops 8 of 14 benchmarks, including MMLU-Pro, SuperGPQA, BBH and GSM8K. And it remains competitive with Qwen3.7-Plus-Base on the rest. Its 51B N-gram embedding parameters use deterministic lookups, adding no per-token https:/…
At a 1M-token context length, QSA’s attention kernel is up to 7.6× faster in prefill and 4.9× faster in decode. With a 90% prefix-cache hit rate, Qwen3.8-Flash-Next delivers 8.6× the prefill throughput of Qwen3.7-Plus. https://t.co/QZ0koCWvQU
Model Architecture Four core upgrades for maximum capability, efficiency, capacity, and stability: - Attention: GDN + QSA Hybrid. Gated DeltaNet (GDN) compresses history. Qwen Sparse Attention (QSA) uses a lightweight indexer for micro-block context selection. Lower the cost of…
Language performance & Vision Language performance: https://t.co/PkhyKwFmP3
⚡Meet Qwen3.8-Flash, a multimodal MoE and an early preview of the Qwen4 architecture, now open-weight! The production version Qwen3.8-Flash will be available soon via QwenCloud API at just $ 0.16/1M input tokens and $ 0.47/1M output tokens. 125B parameters + 51B N-gram https://…