Thinking Machines Lab, co-founded last year by former OpenAI CTO Mira Murati, announced its first #Artificial Intelligence model called Inkling, while giving a nod to Chinese open-source models.
Thinking Machines said in a blog post that Inkling’s architecture design “largely follows” DeepSeek’s V3 model, which was released in late 2024. For a part of Inkling’s post-training process, the U.S. firm also used data generated by open-source models including Chinese startup Moonshot AI’s Kimi K2.5.
Both OpenAI and Anthropic have repeatedly accused Chinese model developers including DeepSeek of distilling their models, saying the practice violated their usage policy.
For Thinking Machines, Inkling is a modest first step. “Inkling is not the strongest overall model available today, open or closed,” the firm said. “We trained it to be a broad, balanced foundation model: strong across many domains, flexible enough to adapt.”
#🇨🇳 ChinA.I. 🤖🧠🦾🤖
https://thinkingmachines...
