Unverified50% confidenceFactExact time
Microsoft's Phi series and Google's Gemma and other small models make extensive use of distillation techniques, enabling them to deliver strong reasoning performance on edge devices.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/6/2026
First Seen
Valid until: 12/5/2026
Sources
Related Entities
Related Claims
Unverified微软的Phi系列小模型和Google的Gemma都大量使用了合成数据进行训练78% similarUnverifiedGoogle's Flash series models achieve high cost-effectiveness using techniques such as model distillation, Sparse Mixture of Experts architecture (SMoE), and inference-time computation optimization.74% similarUnverifiedGoogle open-sourced the Gemma series of small models, drawing on the successful path of Meta's open release of the LLaMA series.71% similarUnverified谷歌同时掌控芯片、操作系统、应用框架（AICORE）和模型四个层次，理论上可实现最深度端到端优化70% similarUnverifiedMicrosoft的Phi系列和Google的Gemma系列是小语言模型的代表性工作70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/865350API
curl https://kongchang.com/api/v1/knowledge/claims/865350MCP
get_claim(id=865350)