Thinking Machines' Inkling 2 Small heads for Apache 2.0 open-weights release this week
The 594B-parameter MoE model, released on Tinker on 5 November, is expected to open its weights under Apache 2.0 within days, independent trackers suggest.
Thinking Machines has scheduled the open-weights release of Inkling 2 Small, its flagship reasoning model, promising Apache 2.0 licensing after external safety testing concludes. The model became available on the company's Tinker platform on 5 November and is a sparse mixture-of-experts system with 594 billion total parameters and roughly 17 billion active per token, with multimodal input support.
Benchmarks cited by Artificial Analysis place the model at an intelligence index of 31, up 18 points from its predecessor, with hallucination rates falling from 61 percent to 27 percent. Independent open-weights trackers estimate the download could land around 12 November, though no laboratory-confirmed date has been published. The company's pricing on Tinker stands at $0.30 per million input tokens and $1.00 per million output tokens. Known weaknesses include a time-to-first-token of roughly 30 seconds and output token consumption about four times higher than the previous generation.