Models

Moonshot AI open-source MoE model with 1T total parameters, 32B activated parameters, 256K context, and text input on Gonka; focused on long-horizon coding, coding-driven design, and agentic workflows

MiniMax open-weight text-generation model focused on agentic coding, software engineering workflows, log analysis, troubleshooting, refactoring, and complex productivity tasks

Updated Qwen3 MoE FP8 non-thinking instruction model with 235B total parameters, 22B activated parameters, and native 262K context; focused on instruction following, reasoning, coding, tool usage, multilingual and long-context tasks