Unverified50% confidenceFactExact time
Decision 3.0 model weights use BF16 precision, can be loaded via the Transformers library (requiring trust_remote_code=True), and training was completed on AMD Instinct MI325X accelerator cards.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn 8 AMD Instinct MI355X accelerator cards, using the vLLM inference framework, the GLM-5.3 model achieved a single-user decoding speed of 469 tok/s.67% similarUnverifiedThis benchmark was conducted on four model families — Gemma, Qwen, Kimi, and MiniMax — using AMD Instinct MI300X and MI355X hardware.67% similarUnverifiedAMD and EmbeddedLLM jointly published a blog documenting the process of optimizing MiniMax M3 inference serving on Instinct MI355X hardware.64% similarUnverifiedAMD Instinct系列（如MI300X）通过ROCm软件栈实现与PyTorch、TensorFlow等主流AI框架的兼容63% similarVerifiedThe MI355X belongs to AMD's Instinct series of data center accelerator cards and is evolved from the CDNA architecture.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1003461API
curl https://kongchang.com/api/v1/knowledge/claims/1003461MCP
get_claim(id=1003461)