Inkling-Small
Made by Thinking Machines, released 15 Jul 2026. Epoch AI lists its access as “Open weights (unrestricted)”. No Top tier or Strong results yet among the models listed here.
In short
What it does well
No Top tier or Strong ratings on the four main tests yet.
Where it falls short
- Coding: Behind the leaders (70th of 77 listed models)
- Knowledge: Behind the leaders (43rd of 43 listed models)
- Maths: Behind the leaders (36th of 43 listed models)
- Writing: Behind the leaders (77th of 77 listed models)
What we don't know yet
- Fixing real software bugs: Not enough data
- Difficult coding tasks: Not enough data
The four ratings
- Top tier
- Strong
- Middle
- Behind the leaders
- Not enough data
Ratings compare only the 77 models listed here, in quarters, best first. “Not enough data” means no result yet. It is never a low score.
- Writing
- Behind the leadersHow often people preferred its answers to writing prompts. 1,336 on Arena's scale · Source for Inkling-Small writing score (lmarena)
- Coding
- Behind the leadersHow often people preferred its answers to coding questions. 1,451 on Arena's scale · Source for Inkling-Small coding score (lmarena)
- Maths
- Behind the leadersHow many very hard maths problems it solved. 46.3% · Source for Inkling-Small maths score (epoch)
- Knowledge
- Behind the leadersHow many short factual questions it answered correctly. 19.1% · Source for Inkling-Small knowledge score (epoch)
Evidence completeness: Broad evidence, 11 of 13 tests have a result.
Where it is relatively stronger or weaker
This compares each kind of request with the model's own overall position among the 77 listed models, using Arena's votes. It is about performance on different languages, kinds of request and topics, not about political or social bias.
Relatively stronger
No category stands out as clearly stronger than its overall position.
Relatively weaker
No category stands out as clearly weaker than its overall position.
Show every category we can see
| Category | Position | Arena score | Votes | Source |
|---|---|---|---|---|
| Expert-level prompts Kind of request | 66th of 77 | 1,441 95% range 1,431 to 1,450 | 4,292 | Source for Expert-level prompts |
| Hard prompts Kind of request | 71st of 77 | 1,424 95% range 1,419 to 1,429 | 24,140 | Source for Hard prompts |
| Long prompts Kind of request | 74th of 77 | 1,404 95% range 1,398 to 1,410 | 17,285 | Source for Long prompts |
| Longer conversations Kind of request | 70th of 77 | 1,416 95% range 1,407 to 1,425 | 5,706 | Source for Longer conversations |
| Chinese Language | 68th of 76 | 1,464 95% range 1,453 to 1,476 | 2,975 | Source for Chinese |
| English prompts Language | 70th of 77 | 1,420 95% range 1,414 to 1,426 | 14,988 | Source for English prompts |
| French Language | 64th of 72 | 1,435 95% range 1,417 to 1,453 | 1,235 | Source for French |
| German Language | 52nd of 55 | 1,404 95% range 1,379 to 1,428 | 626 | Source for German |
| Japanese Language | 45th of 54 | 1,394 95% range 1,368 to 1,419 | 626 | Source for Japanese |
| Korean Language | 51st of 54 | 1,374 95% range 1,350 to 1,398 | 658 | Source for Korean |
| Non-English prompts Language | 71st of 77 | 1,403 95% range 1,397 to 1,408 | 22,004 | Source for Non-English prompts |
| Russian Language | 73rd of 77 | 1,396 95% range 1,387 to 1,406 | 4,500 | Source for Russian |
| Spanish Language | 62nd of 70 | 1,427 95% range 1,409 to 1,445 | 1,134 | Source for Spanish |
| Entertainment, sport and media Topic | 75th of 77 | 1,353 95% range 1,345 to 1,360 | 9,634 | Source for Entertainment, sport and media |
| Law and government Topic | 71st of 76 | 1,414 95% range 1,403 to 1,425 | 3,157 | Source for Law and government |
| Medicine and health Topic | 67th of 75 | 1,414 95% range 1,402 to 1,427 | 2,650 | Source for Medicine and health |
| Science Topic | 71st of 77 | 1,427 95% range 1,419 to 1,436 | 6,110 | Source for Science |
| Software and IT Topic | 71st of 77 | 1,448 95% range 1,442 to 1,454 | 14,266 | Source for Software and IT |
Bias: what we can and can't see
Our sources do not measure political lean, stereotyping, flattery (agreeing with you too readily) or safety behaviour, so we cannot say how any model does on those. What we can show is above: where a model is relatively stronger or weaker across languages, kinds of request and topics. Also, people voting in Arena can favour answer style, such as length and formatting, and our numbers do not adjust for that. See what AI bias means and what these tests can tell you.
Best-supported use cases
This model is not among the options we recommend for any of the five tasks. That reflects the rules and the models listed, not a verdict on the model.
Where an alternative might be more suitable
- Writing and creating content: rated Behind the leaders on writing votes. Compare it with Claude Opus 5.5, the best-supported choice for this task.
- Coding and building things: rated Behind the leaders on coding votes. Compare it with Claude Opus 5.5, the best-supported choice for this task.
- Homework and maths: rated Behind the leaders on hard maths problems. Compare it with GPT-6.1 Sol, the best-supported choice for this task.
- Business emails and everyday work: rated Behind the leaders on business-task votes. Compare it with Gemini 4 Argon, the best-supported choice for this task.
- Research and checking facts: rated Behind the leaders on short factual questions. Compare it with Gemini 4 Argon, the best-supported choice for this task.
What it is
- Maker
- Thinking Machines
- Release date (Epoch AI)
- 15 Jul 2026
- Access, as listed by Epoch AI
- Open weights (unrestricted)
- Licence, as listed by Arena
- Apache 2.0
- Maker's country (Epoch AI)
- United States of America
- Last observation in our data
- 28 Oct 2026
- Data last changed
- 29 Oct 2026, 14:36 UTC
We have not verified an official product page for this model, so we do not link to one.
How you can use it
Open weights. Arena lists its licence as Apache 2.0. In plain terms, the model's files are published under a licence that lets others run it, so it can be offered by more than one company, sometimes at lower cost. That does not guarantee that any particular app is free. Open licences differ, and some restrict commercial use, so check the licence terms before building a business on it.
How you can reach it (an app, a website, a developer API or a third-party tool) and what that costs are not covered by our benchmark data. Where we could verify prices from the maker's own pages, they are shown under Price. Epoch AI lists access as “Open weights (unrestricted)”. See models, apps and harnesses and how AI pricing works.
Price
Price not independently verified
We have not been able to check this maker's own pricing page, so we show nothing for it.
We never infer subscription prices from API prices. New to “tokens” and “per million”? See how AI pricing works.
Good to remember
- Benchmarks are samples of tasks. None proves performance in every situation.
- Arena scores come from people choosing between answers. That measures preference, not correctness.
- Ratings are relative to the 77 models listed here. A different set would change them.
- Sources test models at different settings. Each result shows the setting used, and we do not adjust it.
The evidence, for the curious
Show all 13 tests with settings and sources
| Test | Result | Rating | Setting used | Source identifier | Observed | Links |
|---|---|---|---|---|---|---|
| Arena text: overall preference Arena 'text' leaderboard (Bradley-Terry rating, no style control) |
1,414 Arena score 95% range 1,410 to 1,419, 36,999 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Inkling-Small lmarena · text/latest |
28 Oct 2026 | Source for Inkling-Small Arena text: overall preference score (lmarena) |
| Arena text: creative writing Arena 'text' leaderboard, category creative_writing |
1,336 Arena score 95% range 1,328 to 1,344, 7,972 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Inkling-Small lmarena · text/latest |
28 Oct 2026 | Source for Inkling-Small Arena text: creative writing score (lmarena) |
| Arena text: coding Arena 'text' leaderboard, category coding |
1,451 Arena score 95% range 1,444 to 1,458, 10,049 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Inkling-Small lmarena · text/latest |
28 Oct 2026 | Source for Inkling-Small Arena text: coding score (lmarena) |
| Arena text: maths Arena 'text' leaderboard, category math |
1,458 Arena score 95% range 1,444 to 1,473, 1,727 votes |
Middle | effort: not stated, bradley-terry rating | Inkling-Small lmarena · text/latest |
28 Oct 2026 | Source for Inkling-Small Arena text: maths score (lmarena) |
| Arena text: following instructions Arena 'text' leaderboard, category instruction_following |
1,401 Arena score 95% range 1,395 to 1,407, 13,454 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Inkling-Small lmarena · text/latest |
28 Oct 2026 | Source for Inkling-Small Arena text: following instructions score (lmarena) |
| Arena text: business and management Arena 'text' leaderboard, category industry_business_and_management_and_financial_operations |
1,414 Arena score 95% range 1,406 to 1,422, 6,813 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Inkling-Small lmarena · text/latest |
28 Oct 2026 | Source for Inkling-Small Arena text: business and management score (lmarena) |
| Arena WebDev: building web apps Arena 'webdev' leaderboard (Code Arena), category overall |
1,409 Arena score 95% range 1,400 to 1,418, 5,295 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Inkling-Small lmarena · webdev/latest |
28 Oct 2026 | Source for Inkling-Small Arena WebDev: building web apps score (lmarena) |
| FrontierMath Tiers 1-3 (v2) FrontierMath Tiers 1-3 v2 (private set), benchmark released 2026-06-12 |
46.3% share of problems solved standard error ±3.0 points |
Behind the leaders | effort: xhigh, best score across scorers | inkling-small_xhigh epoch · frontiermath_tiers_1_3_v2.csv |
14 Aug 2026 | Source for Inkling-Small FrontierMath Tiers 1-3 (v2) score (epoch) · Run details for FrontierMath Tiers 1-3 (v2) |
| OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025, benchmark released 2024-12-19 |
90.0% share of problems solved standard error ±4.0 points |
Middle | effort: xhigh, best score across scorers | inkling-small_xhigh epoch · otis_mock_aime_2024_2025.csv |
14 Aug 2026 | Source for Inkling-Small OTIS Mock AIME 2024-2025 score (epoch) No public run log in the data file. |
| GPQA Diamond GPQA Diamond, benchmark released 2023-11-20 |
88.5% share of questions correct standard error ±1.9 points |
Middle | effort: xhigh, best score across scorers | inkling-small_xhigh epoch · gpqa_diamond.csv |
14 Aug 2026 | Source for Inkling-Small GPQA Diamond score (epoch) No public run log in the data file. |
| SimpleQA Verified SimpleQA Verified, benchmark released 2025-09-09 |
19.1% share of questions correct standard error ±1.2 points |
Behind the leaders | effort: xhigh, best score across scorers | inkling-small_xhigh epoch · simpleqa_verified.csv |
27 Aug 2026 | Source for Inkling-Small SimpleQA Verified score (epoch) · Run details for SimpleQA Verified |
| SWE-bench Verified (Epoch runs) | Not enough data | |||||
| FrontierCode (reported by Cognition) | Not enough data | |||||
Missing evidence
Not enough data for these tests:
- SWE-bench Verified (Epoch runs)
- FrontierCode (reported by Cognition)
Missing results are not low scores and do not count against the model.
Where a source gives no public run log, we cannot show the run behind the score. See data sources.
Not sure this is the right one?
Tell us what you need help with and we'll narrow the options down.