Unverified50% confidenceFactExact time
Some of DeepSeek's models restrict using their outputs to train competing models.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
UnverifiedDeepSeek将训练成本压缩至业界同类模型的十分之一左右75% similarUnverifiedDeepSeek开放模型存在token效率偏低的问题，开源和部分国产模型在此方面普遍弱于Claude和GPT系列73% similarUnverifiedThe method is training-free, low-cost, and diagnosable, and can be plug-and-play with any open-source LLM, but is strongly model-dependent and will fail on models such as DeepSeek.72% similarUnverifiedThe training cost of DeepSeek R1 has been compressed to a fraction of that of comparable models.72% similarVerifiedDeepSeek has chosen to open-source its model weights.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1004628API
curl https://kongchang.com/api/v1/knowledge/claims/1004628MCP
get_claim(id=1004628)