Research·Global

Thinking Machines Unveils 975B Parameter Model, Highlights US-China AI

Global AI Watch · Editorial Team··5 min read
Thinking Machines Unveils 975B Parameter Model, Highlights US-China AI
Editorial Insight

Inkling is now the largest US open-weights model but trails in performance against top Chinese AI models.

Key Points

  • 1Third largest open-weights AI model globally, behind China.
  • 2Advances multi-modal capabilities over prior US models.
  • 3Increases reliance on large-scale, open-weight models.

What Changed

Thinking Machines Lab, established by Mira Murati, has launched Inkling, a multimodal model with 975 billion parameters. This places it as the largest U.S. open-weights model to date but remains outperformed by its Chinese counterparts on critical tasks. Notably priced at $1.87 per million input tokens, Inkling is designed for fine-tuning rather than outright computational supremacy. This development highlights a significant move in AI scaling within the United States, though it still trails China in performance benchmarks.

Strategic Implications

The introduction of Inkling underscores a pivotal shift for U.S. AI capabilities. As the largest model of its kind in the U.S., it enhances domestic competition against top-tier Chinese models. However, China's leadership in task performance indicates a persistent edge in AI application efficiency. U.S. labs could gain leverage in areas where large-scale models are essential, but may still face challenges in global competitiveness.

What Happens Next

With Inkling's release, the strategic focus may shift towards closing the performance gap between U.S. and Chinese models. We can expect engagements with U.S. policy makers to enhance support for domestic AI advancements. By mid-2027, the emphasis may include increased government and private sector collaboration to foster technological parity or even leadership.

Second-Order Effects

Inkling's release could induce ripple effects across AI-related industries, including cloud computing and semiconductor manufacturing, as demand for infrastructure to support such models increases. Furthermore, it may prompt regulatory revisions to address the ethical and sovereignty implications of such large-scale AI distributions, affecting both domestic policy and international collaborations.

Free Daily Briefing

Top AI intelligence stories delivered each morning.

Subscribe Free →

Explore Trackers