Thinking Machines Releases Inkling Video‑Audio Model
Originally published Jul 15, 2026
By Will Knight · Wired — AI
AI-generated summary based on Wired — AI · Aggregated by OffScreenSpace · Human-reviewed and approved on Jul 15, 2026
Key points
- Inkling contains 975 billion parameters.
- The model processes video and audio data.
- It is open source and targets competitive positioning.
Thinking Machines has introduced Inkling, an open source model with 975 billion parameters designed to understand both video and audio content. The release aims to position the company alongside rivals such as Anthropic and OpenAI.
Read the original story: Wired — AI — by Will Knight