Qwen3.6 27B
Open weightsGeneral LLMAudioReasoningTool useVision
256Kcontext
$0.6 / $3.6per 1M tokens (in / out)
27.8Bparameters
17stories in our feed
Specs
hover a value to see its sourceContext window256K
Max output64K
Parameters27.8B
Knowledge cutoff—
Licenceapache-2.0
Commercial useCommercial OK
InputText · Image · Video
OutputText
HF downloads (30d)2.3M
Familyqwen
API price
USD per 1M tokensInput$0.6
Output$3.6
Cached input—
No price change recorded since we started tracking.
News timeline
from the AITOP feed, one story per event2026-11-06
MAEM 方法用 RL 生成文本样例,解读 Qwen3.6-27B 残差流方向arXiv cs.LG
2026-11-05
斯坦福研究:NVIDIA B200 每单位能源产出为 AMD MI350X 的 9 倍以上rohanpaul_ai
2026-11-02
2026-10-17
2026-10-05
Figure 将推出主动式 AI 助手 Hark,支持多模型调度与订阅计划TestingCatalog
2026-09-18
2026-09-09
Agent 通过率提升 19%,模型参数未动shao__meng
2026-08-27
LocalLSTC:用于本地GUI代理的长期短期控制架构arXiv cs.AI
2026-08-11
阿里确认Qwen3.8-27B将发布,性能更强阿里通义 Qwen
2026-08-10
Transformer难用涌现世界模型:河内塔任务新研究arXiv: DeepSeek
2026-08-03
2026-07-30
MindForge: 通过源码无关程序合成教小语言模型全生命周期软件工程arXiv cs.LG
2026-07-14
2026-07-10
TRACE:智能体识别缺失能力自我改进,Qwen3.6-27B达73.2%Stanford AI Lab
2026-06-30
笔记本上能跑的本地模型推荐:Qwen3.6-27B等5款Paul Couvert