Intellect-3: A 100B+ MoE trained with large-scale RL

• Posted 7 hours ago by anacleto
• 3 points