Unverified50% confidenceFactExact time
一些推理增强型模型（如DeepSeek-R1、Kimi K3）会生成非常长的思维链，有时推理过程的Token数量是最终答案的数十倍
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/26/2026
First Seen
Valid until: 11/24/2026
Sources
Related Entities
Related Claims
UnverifiedReasoning models, represented by OpenAI's o series, DeepSeek-R1, and the Kimi K series, perform a chain-of-thought process before outputting, decomposing problems into multiple steps, performing self-verification, and deriving answers incrementally.80% similarUnverifiedReasoning models (such as OpenAI's o1/o3 series, Anthropic's Claude with extended thinking, and DeepSeek-R1) output their thinking process (thinking tokens) before generating the final answer.79% similarUnverifiedThe Kimi model occasionally falls into lengthy reasoning token loops when using chain-of-thought reasoning, oscillating back and forth between two mutually contradictory conclusions.78% similarUnverifiedKimi K3属于推理模型类别，通过强化学习训练使模型在生成最终答案前进行显式思维链推演77% similarUnverifiedKimi K3内置常开的思考模式，默认在生成回答前进行推理链条思考76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/804746API
curl https://kongchang.com/api/v1/knowledge/claims/804746MCP
get_claim(id=804746)