Thinking Machines: Inkling
thinkingmachines's mid-tier model. Long-context specialist with 1.0M window.
Scores by test
Methodology →What you need to know
Inkling is an open-weight model optimized for high-reasoning tasks, specifically strategic analysis, agentic planning, and structured output. It achieves perfect internal scores in faithfulness and tabular data processing, making it a reliable choice for complex data extraction and logical synthesis where accuracy is non-negotiable. Its 1.0M token context window provides significant headroom for processing large documents, though its performance in long-context scenarios is slightly lower than its peak reasoning capabilities.
The model's pricing is positioned in the mid-to-high tier, with a blended cost of $3.29/MTok. While more expensive than many lightweight open-weight alternatives, the cost is justified for developers requiring high-fidelity structured outputs and strategic depth. However, the value proposition drops for simpler tasks; the model struggles with basic classification and constrained rewriting, where cheaper, smaller models often perform better.
Tool calling is a notable limitation, scoring only 3/5, which suggests this model should not be the primary driver for complex API orchestration or autonomous agent loops requiring precise function calling. Its strength lies in the cognitive processing of information rather than the execution of external commands.
Use this model if your workflow requires high-faithfulness strategic planning, complex tabular data analysis, or multilingual structured outputs. Skip this model if your primary use case is simple text classification, strict constrained rewriting, or heavy reliance on tool calling.
Strengths — Top 3
Relative weaknesses — Bottom 3
Similar models