Records found: 3
'A concise technical comparison of vLLM, TensorRT-LLM, Hugging Face TGI v3 and LMDeploy, highlighting when to use each stack for production LLM inference based on throughput, latency and KV behavior.'
'MiniMax-M2 is a compact MoE released with open weights on Hugging Face, optimized for coding and agentic tool use with about 10B active parameters per token and claims of ~2x speed at ~8% of Claude Sonnet cost.'
Alibaba's Qwen team published FP8-quantized checkpoints for Qwen3-Next-80B-A3B (Instruct and Thinking), enabling more efficient long-context inference and practical deployment of an 80B/3B-active hybrid MoE on commodity GPUs