Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, released its first open-source large language model, Inkling, on July 16. The model boasts 975 billion total parameters, is fully open-weight under the Apache 2.0 license, and matches the performance of Nvidia's flagship Nemotron 3 Ultra in authoritative benchmarks while using only one-third of the inference tokens. Inkling emphasizes native multimodality and extreme customizability, deeply integrated into the company's Tinker fine-tuning platform, aiming to capture the enterprise AI infrastructure market. Meanwhile, Chinese AI startup Baichuan Intelligence saw its last co-founder depart after founder Wang Xiaochuan insisted on pivoting from general-purpose models to AI healthcare, creating a stark contrast with Inkling's open ecosystem and highlighting the ongoing industry debate between model capability and vertical applications.