Expired50% confidenceFactExact time
Zhipu AI stated that GLM-5.2 was trained on full-stack domestic chips.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
GLM-5.2 Deep Dive: Million-Token Context, MIT Open Source & Full-Stack Domestic Chip Training
bilibiliAI技术投降派6/14/2026
Related Claims
UnverifiedGLM-5 and GLM-5.1 have completed inference adaptation for 7 domestic chips: Huawei Ascend, Moore Threads, Cambricon, Kunlunxin, Metax, Enflame, and Hygon.69% similarUnverifiedThe GLM-5 series was trained end-to-end on Huawei Ascend 910B chips, using approximately 100,000 units, paired with Huawei's self-developed MindSpore framework, with zero dependence on NVIDIA throughout.69% similarUnverifiedggml针对ARM NEON、AVX2、AVX-512等SIMD指令集做了专项优化63% similarUnverifiedGGML是Gerganov自研的张量运算库，专为CPU推理优化，采用纯C实现，可在ARM、x86等架构上运行62% similarUnverifiedNVIDIA offers a complete product line ranging from high-end training chips (H100/B200) to low-cost inference chips.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/45332API
curl https://kongchang.com/api/v1/knowledge/claims/45332MCP
get_claim(id=45332)