AI Matters

Why it matters

Thinking Machines releases Inkling-Small multimodal model

MarkTechPost Thinking Machines Lab Releases Inkling-Small: A 276B Total, 12B Active Open Weights Multimodal MoE Model

Thinking Machines Lab released Inkling-Small, a 276-billion parameter mixture-of-experts model with 12 billion active parameters that runs on one Nvidia GPU.

Why it matters

The open-weights model matches the performance of larger versions while running efficiently on a single graphics card, lowering hardware costs.