Back to Models
Thinking Machines: Inkling Small
Thinkingmachinestext8.8 / 10 Overall Rating
Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines with 12B active out of 276B total parameters. It features a 1M token context window and supports text, image, and audio inputs at a fraction of standard API costs.
Context Window
1049K
Knowledge Cutoff
2025
Max Output
16K
License
Proprietary
Where it Excels (Pros)
- 1M token context window
- Significantly lower API pricing
- Native audio and vision inputs
- Massive 262K max output limit
Limitations (Cons)
- High VRAM requirement for hosting
- 12B active params limits reasoning
Benchmark Breakdown
Reasoning (MMLU)82%
Coding (HumanEval)78%
Mathematics (MATH)76%
Long-Context Retrieval95%
Pricing Matrix
Input cost / 1M tokens$0.45
Output cost / 1M tokens$1.20
Compare Thinking Machines: Inkling Small with other frontier models: