Models

Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use

Lightly post-trained 398B MoE chat model for creative work, long-context prompts, and tool-using agents

Experimental chat-tuned 6B MoE model with 1B active parameters for low-resource chat and instruction following

Reasoning-tuned 26B MoE model with 3B active parameters for agents, tools, and multi-step workloads