GoKawiilNew AI lab releases Inkling 2 Small, a 594B-parameter mixture-of-experts model
The developer unveiled Inkling 2 Small, a Mixture-of-Experts model with 594 billion total parameters and 17 billion active, supporting image inputs and variable effort levels. The model scored 31 on the Artificial Analysis Intelligence Index, a notable jump from the company's prior generation, with gains attributed to more efficient pre-training and a unified reinforcement learning run. It is currently accessible only through the Tinker platform with closed weights, ahead of a planned Apache 2.0 open-weight release next week pending safety testing.
GoKawiil's interpretation of the reporting above, not reported fact.
The staged rollout—closed weights first, open weights after external red-teaming—reflects a cautious approach to releasing increasingly capable models, which the company frames as a 'safe path to open weights.' Strong performance in agentic coding and long-horizon tasks, demonstrated through examples like a procedurally generated city simulator and a from-scratch Linux PC emulator, suggests the model could be aimed at developers building complex autonomous agents. The eventual open-weight release could broaden access to a highly capable model, intensifying competition among providers of large open models.
- Inkling 2 Small has 594B total parameters with 17B active, and supports image inputs and variable effort.
- It scored 31 on the Artificial Analysis Intelligence Index, surpassing the developer's prior generation.
- Full weights are expected under an Apache 2.0 license next week, pending completion of safety testing and red-teaming.
Source: thinkingmachines.ai — Thinking Machines Lab, 2026-11-06
Published there as: “Inkling 2 Small”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.