Thinking Machines

Thinking Machines Lab describes itself as an artificial intelligence research and product company, based in San Francisco.^{1, 4, 5, 6}

It was publicly launched in February 2025 by Mira Murati, a former chief technology officer of OpenAI.^{2, 3, 7, 8}

The company says it aims to make AI systems more widely understood, customizable and generally capable.¹

On 1 October 2025 it launched Tinker, an API for fine-tuning language models.^16, 17

Tinker lets users fine-tune open-weight models, including Qwen-235B-A22B, and Inkling is also offered for fine-tuning on the platform.^9, 16

On 11 May 2026 the company announced a research preview of interaction models, which take in audio, video and text continuously and respond in real time.^18, 19

On 15 July 2026 it released Inkling, its first in-house model, with open weights under the Apache 2.0 license.^{9, 10, 11, 12, 13}

Inkling is a mixture-of-experts transformer with 975 billion total parameters, of which 41 billion are active, and a context window of up to 1 million tokens.⁹

Inkling-Small, released on 30 July 2026, is a smaller open-weights model with 276 billion total parameters, 12 billion of them active.^13, 20

The company states its mission as building AI that extends human will and judgment.¹⁵

As of March 2026, Mira Murati is the cofounder and CEO of Thinking Machines Lab.^2, 3