Inkling 2 Mini runs locally, rivals larger AI models

Inkling 2 Mini runs locally, rivals larger AI models
Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, has released Inkling 2 Mini, an open-weights multimodal AI model that runs offline on a 32GB MacBook Pro or single desktop GPU. At 26.4 billion parameters with a roughly 20GB footprint, it uses a looped transformer architecture that cycles data through middle layers three times per token, yielding 48 effective layer applications without increasing storage size. The model ties Alibaba's dense 27B flagship on intelligence benchmarks while running four times faster and six times cheaper per task.
Read the original article →