Inkling Small
Thinking Machines · released Jul 27, 2026 · thinkingmachines/Inkling-Small
- Type
- Open weightsApache License 2.0
- Params
- 266B
- Context
- 524K
about 393K words of context
Our take
Written Sep 17, 2026Inkling Small is a downloadable model that takes text, images and audio and returns text, under a licence allowing commercial use, changes and redistribution. Its measured quality comes from human-preference arenas rather than correctness tests, and on the agent board its task-outcome and steerability results sit below the board's neutral point.
Use it for everyday chat, writing and question-answering where a downloadable model with a permissive licence matters, or when a screenshot or a voice clip arrives alongside a written question. Skip it if you need an agent that reliably finishes a multi-step task on its own, or if you need measured correctness on coding and maths.
The case for it
- The licence allows commercial use, changes and redistribution (Apache License 2.0).
- Text, images and audio all go into the same request, so a screenshot or a recording does not have to be transcribed or described first.
- The request capacity takes a long report or a stack of documents without splitting them up first, though reliable recall across all of it is unverified in our data.
- Coding prompts are among its stronger arena results, from human pairwise votes rather than a correctness rubric.
The case against it
- On the agent arena its task-outcome score sits below the board's neutral point, so it is not the model to hand a multi-step job unsupervised.
- Every score supplied is an arena preference or inverse-propensity figure; nothing measures correctness on a fixed test set, so coding and maths ability need a trial on work you can check yourself.
- No host carries a throughput figure, so the cheapest listed rate cannot be weighed against how fast it answers.
How good is it?
An open text model for everyday writing and questions, though multi-step tasks and tool calls are not its strength.
- carrying out multi-step tasks for youArena Agent · 46th of 55
- calling tools to carry out requestsArena Agent · Tool use · 44th of 55
- changing course when you give new instructionsArena Agent · Steerability · 54th of 55
EverydayGeneral questions and everyday reasoning
Arena Text (overall)93rd of 168 · 1405
CodingWriting and fixing code on its own
Arena Coding73rd of 168 · 1473
AgenticPlanning, calling tools, staying on task
Arena Agent46th of 55 · −0.096
Arena Agent is the only board that has scored it for this.
WritingDrafting and rewriting prose
Arena Creative Writing120th of 168 · 1313
Arena Creative Writing is the only board that has scored it for this.
Placings on Arena's agent boards, from live sessions people ran themselves. A model can lead on one of these and sit mid-field on the others.
Boards this model appears on that none of the ratings above are built on.
Every published score for this model12 scoresEvery figure we hold, from 12 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Comfortable fit
Apple M3 Ultra (80-core GPU) · 512 GB
Room to spare. 209.7 GB spare means a 10% error in the size would not change the answer.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked 4 hours ago — each listing carries its own date.
- per 1M tokens
- $0.45 in / $1.20 out
- Context served
- 524K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouterOpenRouter's own listing | $0.45 / $1.20checked 4 hours ago | 524K | not measured | Unknown | Unknown | Unknown |
| DeepInfrafp8Direct and through OpenRouter | $0.45 / $1.20checked 4 hours ago | 524K262K max reply through OpenRouter | 130 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
DeepInfraAcross the 2 listings we hold: 1 says it does not train on prompts (only through OpenRouter), 0 say they do and 1 does not say. 1 appears in the zero-retention registry we check (only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouterOpenRouter's own listing | ✓ | ✗ | ✗ |
| DeepInfrafp8Direct and through OpenRouter | ✓ | ✗ | ✗ |
Tool calling: 2 of 2 listings say yes. JSON output: 0 of 2 listings say yes, 2 say no. Strict schema: 0 of 2 listings say yes, 2 say no.
When we formed this view
Recent changes
Each date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 1 of 2 listings does not say whether it trains on prompts, and 1 answers only through OpenRouter, not for its own listing.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsApache License 2.0, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Apache License 2.0
Fully permissive: commercial use, redistribution, and derivatives allowed. Requires attribution and a copy of the license. Includes an express patent grant.
Identifiers
- Hugging Face
- thinkingmachines/Inkling-Small
- Architecture
- Dense
- Takes in, gives back
- Text, images and audio in, text out
- Catalogue slug
- thinkingmachines-inkling-small