Artificial intelligence / news brief
Thinking Machines releases Inkling 2 Small through Tinker ahead of planned open weights
- Event
- Published

Thinking Machines released Inkling 2 Small, a 594B-parameter mixture-of-experts model with 17B active parameters, text-and-image input and a 1M-token context window. It is available for inference and fine-tuning through Tinker with closed weights; the lab says Apache 2.0 weights will follow after external safety testing, so that release is not yet complete.
Why it matters
The launch pairs a large sparse multimodal model with immediate fine-tuning access, while making its promised open-weight status contingent on a later release.
Sources
- Introducing Inkling 2 Small (external site)Thinking Machines Lab · primary source ·
- Thinking Machines Inkling 2 Small: the highest scoring open weights model outside of China pending weights release (external site)Artificial Analysis · coverage ·
- Thinking Machines Releases Inkling 2 Small, a 594B MoE With 17B Active (external site)MarkTechPost · coverage ·