The Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, has launched its first open AI model, Inkling.

Today, we are introducing Inkling.

Inkling reasons efficiently across text, image, and audio modalities. We are making the full weights available.https://t.co/Ghebq5mG30

Available today for fine-tuning on Tinker. Play with it in the Inkling Playground. 🧵— Thinking Machines (@thinkymachines) July 15, 2026

This neural network is built on a MoE architecture: only a portion of the network is activated for each request, speeding up inference without sacrificing depth. It has a total of 975 billion parameters, with 41 billion active. The context window is 1 million tokens. The model was trained from scratch on 45 trillion tokens of text, images, audio, and video.

Inkling can reason across multiple modalities but currently only provides text-based responses, including code and structured data. Users can also adjust the balance between speed and quality of results through a controlled reasoning level.

Alongside Inkling, developers introduced a lighter version, Inkling-Small, which has 276 billion parameters, of which 12 billion are active. The full weights for this model have not yet been released, as the company continues testing it.

Strengths of Inkling

The model has shown particularly impressive results in agent tasks. On the MCP Atlas, Inkling scored 74.1%, nearly 30 points higher than Nvidia's Nemotron 3 Ultra, its main Western competitor among open models. On SWE-Bench Verified, it achieved 77.6% compared to 70.7% for Nemotron.

On the FORTRESS Adversarial benchmark, Murati's neural network scored 78%. It is currently the most powerful open model released by a Western company, although it still lags behind Chinese competitors.

On Terminal Bench 2.1, Inkling scored 63.8%, while Z.ai GLM 5.2 achieved 82.7%. The recently released Kimi K3 reached 88.3% and also leads in Humanity's Last Exam.

Thinking Machines positions Inkling as a "universal" model: excelling in various tasks without sacrificing performance in others. The focus is on customization, as the company believes businesses prioritize a flexible model tailored to specific tasks.

It’s worth noting that Murati left OpenAI in September 2024, citing a desire to pursue her own research.