API pricing

# Thinking Machines Lab API pricing

> Thinking Machines Lab API prices run from $0.45 per million input tokens (Inkling-Small) to $1.87 (Inkling). Its highest-ranked model, Inkling-Small, costs $0.45 input and $1.20 output per million tokens.
- Canonical page: https://noometry.com/llm-pricing/thinking-machines
- Last updated: 2026-10-10
- Title: Thinking Machines Lab API Pricing (October 2026): Every Model per 1M Tokens

Thinking Machines Lab API prices run from $0.45 per million input tokens (Inkling-Small) to $1.87 (Inkling). Its highest-ranked model, Inkling-Small, costs $0.45 input and $1.20 output per million tokens.

Last verified October 10, 2026

API prices per million tokens
|  |  |  | Cached input |  |  |  |
| --- | --- | --- | --- | --- | --- | --- |
| [Inkling-Small](https://noometry.com/models/inkling-small) (open weights) | $0.45 | $1.20 | $0.10 | **$0.64** | 524K | 46.5 |
| [Inkling](https://noometry.com/models/inkling) (open weights) | $1.87 | $4.68 | $0.37 | **$2.57** | 66K | 44.1 |

## Prices on other platforms

The same models are often sold through clouds and resellers at different rates.

| Model | Route | Input | Output | Checked |
| --- | --- | --- | --- | --- |
| [Inkling-Small](https://noometry.com/models/inkling-small) | deepinfra | $0.45 | $1.20 | 2026-10-10 |
|  | openrouter | $0.45 | $1.20 | 2026-10-10 |
| [Inkling](https://noometry.com/models/inkling) | deepinfra | $0.95 | $4.05 | 2026-10-10 |
|  | fireworks | $1 | $4.05 | 2026-10-10 |
|  | openrouter | $1 | $4.05 | 2026-10-10 |
|  | thinking-machines | $1.87 | $4.68 | 2026-10-10 |
|  | together | $1 | $4.05 | 2026-10-10 |

## Frequently asked questions

### How much does the Thinking Machines Lab API cost?

Thinking Machines Lab API prices run from $0.45 per million input tokens (Inkling-Small) to $1.87 (Inkling). Its highest-ranked model, Inkling-Small, costs $0.45 input and $1.20 output per million tokens.

### Does Thinking Machines Lab discount cached input?

Yes. Cached input is billed at a lower rate on 2 of the 2 priced models; the cached rate is listed next to each model.