Inkling (free)

Thinking Machines· released 17 Jul 2026

thinkingmachines/inkling:free
Input / 1M tokens
free
Output / 1M tokens
free
Context window1M tokens

Longest answer: 262K tokens

reads textreads imagereads audioopen weightstool callingvisionreasoning

About

Inkling is an open-weight multimodal mixture-of-experts model from Thinking Machines, with 41B active parameters out of 975B total. It is designed for general-purpose reasoning, coding, agentic systems, and tool use. It accepts text, image, and audio inputs and returns text. The model supports vision and tools, has a 1,048,576-token context window, and allows up to 262,144 output tokens. It is listed as free and was released on 2026-07-17. Its model id is thinkingmachines/inkling:free, and its Hugging Face id is thinkingmachines/Inkling. For builders, the combination of open weights, multimodal inputs, and tool support may fit a range of text-output workflows; the stated modality and limits help set expectations before integration.

Who it is for

Teams exploring general-purpose reasoning, coding, or agentic systems may consider Inkling, especially when image or audio inputs and tool use are relevant. Its free listing and open weights may also be useful factors in evaluation.

What is good

  • Listed as free.
  • Open weights are available.
  • Accepts text, image, and audio inputs.
  • Context window is 1,048,576 tokens.
  • Supports tools and vision.

What to know first

  • Outputs text only.
  • Maximum output is 262,144 tokens.

Verdict

Inkling combines open weights, multimodal inputs, and tool support with a 1,048,576-token context window. It is listed as free, while its output is text and capped at 262,144 tokens.

Details

Lab
Thinking Machinesopenrouter.ai · 3 Oct 2026
Context
1,049Kopenrouter.ai · 3 Oct 2026
Input price
Freeopenrouter.ai · 3 Oct 2026
Output price
Freeopenrouter.ai · 3 Oct 2026
Max output
262,144 tokensopenrouter.ai · 3 Oct 2026
Inputs
text, image, audioopenrouter.ai · 3 Oct 2026
Open weights
Yesopenrouter.ai · 3 Oct 2026

More longest-context models

See the list