Unverified85% confidenceFactTime unknown
Google的Gemini 2.0 Flash Thinking和DeepSeek-R1等模型跟进了o1的推理时计算扩展技术路线
1
Sources
85%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
GitHub 8K星：最全LLM学习资源库深度解析
githubWangRongsheng
Related Entities
Related Claims
UnverifiedGemini 2.5 Flash采用了混合专家（Mixture of Experts, MoE）架构，在推理时只激活部分参数76% similarVerifiedGemini 2.0 Flash是Google DeepMind推出的轻量级多模态模型，主打速度与性价比76% similarUnverifiedcompany-research-agent项目采用双模型策略，使用Google Gemini 2.5 Flash处理快速摘要和初步分析，使用OpenAI GPT-5.1承担深度分析和报告生成74% similarUnverifiedOpenAI的o1/o3系列和Google的Gemini Thinking采用类似扩展思考的慢思考机制73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/33728API
curl https://kongchang.com/api/v1/knowledge/claims/33728MCP
get_claim(id=33728)