Unverified50% confidenceFactExact time
DeepSeek以极低的训练成本实现了与顶级闭源模型接近的数学与代码能力，其R1模型在发布时引发全球关注
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/21/2026
First Seen
Valid until: 10/19/2026
Sources
Related Claims
VerifiedDeepSeek由量化私募巨头幻方科技孵化，其R1系列模型以极低训练成本实现接近GPT-4级别的推理能力84% similarUnverifiedDeepSeek-R1在发布时以远低于同类闭源模型的训练成本，在多个推理基准上与OpenAI o1持平84% similarUnverifiedDeepSeek-R1的推理能力在多个基准测试中表现惊艳，训练成本远低于国际竞品83% similarUnverifiedDeepSeek R1采取开源策略，开发者可通过LoRA、QLoRA等参数高效微调技术以较低算力成本获得垂直领域专属模型81% similarUnverifiedDeepSeek R1等开源LLM在大规模互联网数据上进行预训练，具备强大的通用能力80% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/576843API
curl https://kongchang.com/api/v1/knowledge/claims/576843MCP
get_claim(id=576843)