Inkling

Facts

Noted: released · 2026-07-16
Noted: open-weights AI model · 2026-07-16
Noted: 975 billion total parameters · 2026-07-16
Noted: 41 billion effective parameters · 2026-07-16
Noted: Mixture-of-Experts architecture · 2026-07-16
Noted: handles text, images, audio, and video · 2026-07-16
Noted: supports up to 1 million tokens of context · 2026-07-16
Noted: allows users to adjust compute for reasoning · 2026-07-16
Noted: positioned as foundation for fine-tuning · 2026-07-16
Noted: shifts focus from benchmark performance to adaptability · 2026-07-16

Structured graph also available as JSON at /public/entities/inkling. CC BY 4.0.

All coverage

Aug 9

Thinking Machines Lab Releases Inkling-Small at a Quarter the Size

Thinking Machines Lab released Inkling-Small, an open-weight MoE model with 276 billion total and 12 billion active parameters. It delivers performance comparable to its 975-billion-parameter predecessor Inkling while using far fewer computational resources. The model is available on the Tinker service and as BF16 and NVFP4 weights on Hugging Face.

Jul 16

Thinking Machines Lab Releases Inkling, an Open-Weights AI Model for Fine-Tuning

Thinking Machines Lab has released Inkling, an open-weights AI model with 975 billion total parameters and 41 billion effective parameters using a Mixture-of-Experts architecture. The model handles text, images, audio, and video, supports up to 1 million tokens of context, and allows users to adjust the amount of compute used for reasoning. The company positions Inkling as a foundation for fine-tuning to specific business needs rather than a general-purpose model.