Thinking Machines Lab launches 594B-parameter Inkling 2 Small

Thinking Machines Lab has released Inkling 2 Small, the first model in its second Inkling generation. The sparse Mixture-of-Experts system has 594B total parameters, with 17B active per operation, and supports text and image inputs, variable reasoning effort and a 1M-token context. Audio is not supported.

The model scores 31 on the Artificial Analysis Intelligence Index, compared with 13 for Inkling Small. In the company’s benchmark table, it reached 86.3% on SWE-Bench Pro v2 (Hard), 64.7% on PerceptionBench and 71.8% on EnterpriseBench-CoreCraft. Its Terminal-Bench 4.0 result was 13.1%, indicating uneven performance across tasks.

Inkling 2 Small is currently available through Tinker for inference and fine-tuning, including the Tinker Playground. Weights are closed at launch, while an Apache 2.0 release is planned after external safety testing and red-teaming. Artificial Analysis lists pricing at $0.30 per 1M input tokens and $1.00 per 1M output tokens. Self-hosting hardware requirements have not been disclosed.