Unverified50% confidenceFactExact time
Perplexity does not train its own foundational large language models; instead, it acts as an intelligent routing layer integrating multiple models from companies such as OpenAI, Anthropic, and Meta, and layers on real-time web search capabilities.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/3/2026
First Seen
Valid until: 12/2/2026
Sources
Related Entities
Related Claims
UnverifiedPerplexity 的创始人 Aravind Srinivas 将答案引擎定义为搜索引擎与大语言模型的融合产物70% similarUnverifiedOpenAI、DeepMind等机构的多项研究都证实了大语言模型谄媚问题的普遍性69% similarUnverifiedLarge language models have two inherent weaknesses: their knowledge is cut off at training time and they cannot obtain the latest information, and they can only generate text and cannot directly perform actions.67% similarUnverifiedOpenAI研究员Ilya Sutskever曾提出，纯文本训练的语言模型本质上是在学习世界的压缩表征65% similarUnverifiedModern multilingual large models establish language-agnostic semantic representations during the pre-training phase, building cross-lingual semantic alignment in high-dimensional vector space through joint training on massive multilingual corpora.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/848740API
curl https://kongchang.com/api/v1/knowledge/claims/848740MCP
get_claim(id=848740)