Thinking Machines Lab, an AI startup led by former OpenAI CTO Mira Murati, has released Inkling 2 Small. Third-party tester Artificial Analysis says the model could challenge Chinese models for the lead among open-weight systems. The company said it will release the model's weights next week. Until then, the model is proprietary.
Thinking Machines' model has 594 billion parameters. It accepts image inputs and outputs only text. The company said the model is stronger at coding and easier to use inside agent loops.
Artificial Analysis scored Inkling 2 Small at 31 on its Intelligence Index. That score is above average for its class and ties with Gemini 3.8 Flash. The index tops out at 60 for Gemini 4 Carbon and 59 for Claude Opus 5.5. Artificial Analysis said that if the weights are released as planned, the model would be the highest-scoring open-weight model developed outside China based on current rankings.
The reviewer called the model amongst the leaders in intelligence and reasonably priced for its class. It also said the model is notably slow per task and very verbose. It described the price of $0.30 per million input tokens as somewhat expensive, compared with a class median of $0.24.
Artificial Analysis reported gains over the predecessor. The model improved 41% in automation capabilities and 20% in coding abilities. On the AA-Omniscience v1.1 test, its hallucination rate fell from 61% to 27%, its accuracy rose from 32% to 34%, and its attempt rate fell from 77% to 53%.