Welcome Inkling by Thinking Machines

Hugging Face
Read the full article
Thinking Machines released Inkling, a 1 trillion parameter multimodal large language model that processes image, text, and audio inputs natively. It supports a 1 million token context window, uses a mixture-of-experts architecture with 41 billion active parameters, and is designed for multimodal reasoning and domain adaptation.

More on this story


More in LLM & Text Generation

Researchers from Stanford and UC Berkeley Introduce TokenGate: Token-Gated Attention Makes Long-Context Prefill Up to 11.3x Faster

MarkTechPost

Moonshot releases Kimi K3.1 Pro and says it cut its hacking skills

Covered by 4 sources