Aethir launches its self-developed open-source LLM API platform, Aethir Mesh; top-tier models including DeepSeek V4 and Kimi K2.6 run inference end-to-end on Aethir GPUs.
According to official announcements, Aethir—a decentralized GPU cloud computing infrastructure platform—has officially launched Aethir Mesh, an in-house open-source large language model (LLM) API platform built atop Aethir’s decentralized GPU infrastructure. The platform is now live and offers a unified API endpoint for direct access to leading open-source models, including DeepSeek V4, Kimi K2.6, GLM-5.1, MiniMax-M2.5, and Qwen3.6-27B.