Unverified90% confidenceFactTime unknown
Composer 2.5训练中85%的计算资源被投入到额外的持续预训练和强化学习后训练中
1
Sources
90%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026
Sources
Cursor Composer 2.5深度解析：1/10成本比肩Claude Opus 4.7的编码能力
bilibili多模态实验室
Related Entities
Related Claims
UnverifiedComposer 2采用了双阶段推进策略：中期训练加大规模强化学习78% similarUnverifiedCursor把85%的总计算量投入到在Kimi K2.5基础之上的额外训练和强化学习74% similarUnverified混合精度训练通常可将训练速度提升1.5-2倍并减少近一半的显存占用68% similarUnverified前沿模型训练对算力的需求每年增长约 4-5 倍66% similarUnverified测评项目越全面(基因检测+代谢+骨密度+肌电)综合成本越高但一次性获取信息多，长期训练者更适合分阶段测评(基础体测+专项测评分开做)以匹配训练周期65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/12611API
curl https://kongchang.com/api/v1/knowledge/claims/12611MCP
get_claim(id=12611)