Models Model details

Thinking Machines

Thinking Machines: Inkling Small

thinkingmachines/inkling-small

Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of the Inkling family and is suited for reasoning, coding, agentic workflows, retrieval-augmented generation, instruction following, and multilingual conversation.

Input
$0.45

/ 1M tokens

Output
$1.2

/ 1M tokens

Context window
524.29K

tokens

Limits

Context

524,288

tokens

Max output

262,144

50% of the context window

Parameters

265,956,439,090

Model facts

Author

Thinking Machines

Tokenizer

Other

Modality

Text + Image + Audio -> Text

License

Apache-2.0

Created

Jul 30, 2026

Hugging Face

thinkingmachines/Inkling-Small

Reasoning

Default HighMaxHighMediumLowMinimalNone

Publication status

Hugging Face: verifiedLicense: publishedParameter count: published

Modalities

Input · TextInput · ImageInput · AudioOutput · Text

Generation controls

Frequency penaltyInclude reasoningLogit biasMax tokensMin pPresence penaltyReasoningReasoning effortRepetition penaltySeedStopTemperatureTool choiceToolsTop kTop p

Capability scores

23.5
Agentic index

Better than 57% of models compared

52.9
Coding index

Better than 55% of models compared

25.7
Intelligence index

Better than 63% of models compared