Unverified50% confidenceFactExact time
Qwen3.5 implements a reinforcement learning framework with separated training and inference, capable of supporting a replay system with tens of trillions of tokens.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/29/2026
First Seen
Valid until: 11/27/2026
Sources
Related Entities
Related Claims
UnverifiedThe training of Qwen3.7 Max adopted an environment-driven tool training method, which is a specific application form of reinforcement learning in large model training75% similarUnverifiedQwen 2.5系列训练数据超过18万亿Token，采用Transformer解码器架构并引入分组查询注意力（GQA）机制73% similarUnverifiedQwen2采用Yarn外推与双块注意力（DCA）组合方案提升长序列能力，后训练采用离线DPO加在线DPO的组合65% similarUnverifiedOrnith 是在千问 3.5（Qwen 3.5）基础上进行深度定向强化训练的产物63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/821110API
curl https://kongchang.com/api/v1/knowledge/claims/821110MCP
get_claim(id=821110)