The Thinking Machines Inkling AI model has officially launched as the first general-purpose AI model from the startup led by former OpenAI Chief Technology Officer Mira Murati. The model, called Inkling, is open-weight and built for developers who want to customize and self-host rather than rely on closed systems.
Table of Contents
- Thinking Machines Inkling AI Model: Overview
- Thinking Machines Inkling AI Model: Key Specifications and Capabilities
- What Is Tinker?
- Why This Launch Matters
- Conclusion
Thinking Machines Inkling AI Model: Overview
Thinking Machines Lab has released Inkling, its first AI model, giving developers and companies the ability to download the underlying weights, run the system independently, and fine-tune it for specific tasks.
Unlike closed proprietary systems, this open-weight approach hands more control to the end user. Consequently, the Thinking Machines Inkling AI model is being positioned as a flexible alternative rather than a pure benchmark leader.
- Natively multimodal: handles text, images, audio, and video
- Approximately 975 billion total parameters
- Around 41 billion parameters activated per task
- Trained on roughly 45 trillion tokens
Thinking Machines Inkling AI Model: Key Specifications and Capabilities
The company describes Inkling as strong across many domains and flexible enough to adapt to different use cases. At the same time, Thinking Machines has openly stated that Inkling is not the strongest overall model available today, whether open or closed.
Instead, the emphasis falls on three areas:
- Cost-effectiveness for enterprise deployment
- Customizability through fine-tuning
- Practical, real-world usability over leaderboard rankings
Early commentary has also highlighted competitive agent and coding performance relative to token usage, suggesting the model can be efficient in production settings.
Thinking Machines Inkling AI Model: Where to Access It
- Available for download on Hugging Face
- Accessible through select technology partners
- Released under an open license
- Ready for immediate developer experimentation
What Is Tinker?
Alongside Inkling, Thinking Machines has introduced Tinker, its fine-tuning platform. Tinker is designed to let developers adapt the model for narrower or more specialized tasks without starting from scratch.
This combination of an open-weight base model and a dedicated fine-tuning platform reflects the company’s broader bet: many enterprises will prefer adaptable, self-hostable models over closed systems from the largest AI labs. Readers can learn more directly from the company at Thinking Machines Lab’s official site.
Thinking Machines Inkling AI Model: Why This Launch Matters
The release positions Thinking Machines as a Western alternative in the open-weight space, an area that has recently been dominated by Chinese AI labs. Murati founded the San Francisco-based company after leaving OpenAI, assembling a team with experience from ChatGPT, Character.ai, PyTorch, and other major projects.
The launch also arrives amid intensifying competition across both closed and open AI ecosystems. Thinking Machines has raised substantial capital and continues to focus on research-driven products that prioritize controllability and efficiency.
Looking ahead, the company is expected to release further technical details and benchmarks in the coming days. For ongoing coverage of major AI product launches.
Conclusion
The Thinking Machines Inkling AI model marks an important milestone for Mira Murati’s startup, offering an open, multimodal alternative built around flexibility rather than raw benchmark dominance. As developers begin testing Inkling and Tinker, the coming weeks should reveal how well this customization-first strategy competes against larger, closed AI systems. If adoption continues to grow, Inkling could become a strong open-weight option for enterprises seeking greater control, lower deployment costs, and customizable AI solutions.
I am Rukaiya Kadiwala, an experienced News Content Writer with 6+ years of expertise in hospitality, travel, hotel, restaurant, business, and lifestyle news. Skilled in writing, research, fact-checking, headline creation, and digital publishing, I create accurate, engaging, and high-quality content that informs and attracts readers worldwide.
