Unverified60% confidenceFactExact time
DeepSeek's previous models were trained using Nvidia H800 chips before export restrictions were implemented.
2
Sources
60%
Confidence
Long-term
Relevance
7/24/2026
First Seen
Sources
中国拟限制顶尖AI模型出海，DeepSeek同步布局自研推理芯片
bilibiliAI-seeker7/7/2026
Related Claims
UnverifiedDeepSeek早期模型使用英伟达H800芯片训练完成，H800是英伟达专为中国市场削减互联带宽后推出的H100替代版本83% similarUnverifiedDeepSeek的R1模型是在美国禁售前用英伟达H800芯片训练的83% similarUnverifiedDeepSeek already proposed during the V3.1 period that UE8M0 FP8 was designed for next-generation domestic chips; this time it is fully implemented in DeepInfer, meaning chips that do not support this format will directly suffer in precision or performance.71% similarUnverified2026年8月13日，DeepSeek在MIT许可证下发布了DeepSeek Harness（命令名dsh），模块化程度极高，模型、工具、沙箱乃至工作循环都可以作为插件替换。67% similarUnverifiedDeepSeek 1017 has been optimized for DeepSeek Harness 0.1+ and Tencent CodeBuddy, and the technical report states that the model and the framework will co-evolve, with the capability unit no longer being just the model but the model plus the framework.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/606627API
curl https://kongchang.com/api/v1/knowledge/claims/606627MCP
get_claim(id=606627)