ai Inkling
Other · open weights · writing benchmark
The short version
Inkling sits at #65 of 130 on ToneBench, in the middle of the pack, with a writing Elo of 1523 and an overall score of 78.9 out of 100. We measured it by having it write all 9 of our real YouTube scripts, five times each, then scoring every draft blind against our own finished versions. Here is how it shook out.
Within Other it ranks 3rd of 4. Muse Spark 1.1 (thinking) is the family's top writer here, about 152 Elo ahead, so this config trades some quality for whatever it saves you in cost or speed.
Among open-weights models, it comes in 12th of 47. Open weights still trail the closed frontier on pure voice fidelity, and you can see that in the gap at the top.
What it does best is Substance & Value: it ranks 58th on that dimension at 82.4, above the board average. Its softer spot is Length Discipline (66.5, 77th), which is the thing to watch if that metric matters most for your use.
It was uneven across the 9 scripts. Its best run was the founder announcement / personal origin story (84.4) and its weakest was the personal technical walkthrough / agentic workflow case study (71.8), a spread of about 13 points. Worth knowing if your writing skews toward one of those formats.
On cost, it runs about $0.034/task. DeepSeek V4 Flash (native reasoner alias) scores higher for less money, at $0.004/task, so on pure value this config is not the frontier. Still, it beats 10 models that cost noticeably more.
If you're tuning reasoning effort, we also tested Inkling at other settings. None of them beat this config; the closest is Inkling (high) (Elo 1486), so the only question is how much quality you want to trade for cost or speed.
Skill profile
The shape is the story: the further a corner reaches, the stronger that dimension. Biggest edge over the board: Visual Cues (+8.1); furthest behind: Length Discipline (-0.7).
Per-metric scores
Nine writing dimensions, each scored 0–100 and blended by the editorial weight shown. Rank is against all 130 current-ranked models. The thick bar is this model; the thin lines above and below are the current field's best model (darker) and average (lighter) on the same scale. The small ± number is its run-to-run variation.
Per-article scores
The same 9 real scripts every current-ranked model writes, scored individually. Different formats stress different skills.
Consistency: run-to-run variability
Every cell on this page is the mean of 5 independent runs per script; the ± numbers are the run-to-run standard deviation. Overall, Inkling varies by ± 4.5 points between runs versus a board median of ± 3.8, so it is about as repeatable as the typical model on the board.
Its most volatile dimension is Hook (± 5.3 vs a board median of ± 3.0): two runs of the same brief can land visibly different hook scores.
| Metric | This model | Board median | Verdict |
|---|---|---|---|
| Tone & Voice ? | 80.5 ± 2.5 | ± 2.2 | typical |
| Writing Craft ? | 80.7 ± 2.5 | ± 2.0 | typical |
| Substance & Value ? | 82.4 ± 2.3 | ± 2.6 | typical |
| Flow & Emotion ? | 76.3 ± 2.7 | ± 2.6 | typical |
| YouTube Structure ? | 75.3 ± 4.5 | ± 3.6 | typical |
| Hook ? | 82.6 ± 5.3 | ± 3.0 | swingier than most |
| Length Discipline ? | 66.5 ± 12.6 | ± 8.6 | typical |
| Anti-Slop ? | 83.6 ± 2.6 | ± 2.6 | typical |
| Visual Cues ? | 82.6 ± 5.5 | ± 4.7 | typical |
Measured, not modeled
Numbers we log on every run and rarely talk about. None of these affect the writing scores; cost and latency are informational.
What each judge scored it
The published overall of 78.9 is the consensus of three family-disjoint judges scoring the same 45 stored drafts blind with the identical rubric. The highest and lowest judge differ by 9.1 points on its overall. A typical amount of judge disagreement for this board. How the panel works: methodology.
How we ran Inkling
- Accessed viaTogether AI (serverless) · model page ↗
- Exact model / route id
thinkingmachines/Inkling(provider default reasoning) - Generations5 independent runs × 9 scripts = 45 drafts, stored immutably after generation
- Run dates2026-07-29 → 2026-08-11
- Judges3-judge panel: Claude Opus 5, high effort (Anthropic) + DeepSeek V4 Flash (OpenRouter) + GPT-5.6 Sol, medium reasoning (Codex CLI): every draft scored independently by all three with the identical rubric, blind to model identity, evidence quotes required before each metric score; published score = per-metric consensus mean, verified with a Bradley-Terry refit (methodology)
- Samplingprovider-default sampling, no fine-tuning, no system-prompt changes; every current-ranked model gets the identical brief, style contract and research packet
- What stays privatethe exact rubric and any non-public reference material; source-publication timing is disclosed for exceptions (why)
Frequently asked questions
How good is Inkling at writing?
On ToneBench it ranks #65 of 130 with a writing Elo of 1523 and an overall score of 78.9/100. That score comes from writing our 9 real YouTube scripts five times each and scoring every draft blind against our own finished versions across nine writing dimensions.
Is Inkling the best Other model for writing?
Not quite. Within Other it ranks 3rd of 4; Muse Spark 1.1 (thinking) is the family's best writer here.
Is Inkling good value for the money?
It costs about $0.034/task. DeepSeek V4 Flash (native reasoner alias) scores higher for less, so it is not the value pick.
What are Inkling's strengths and weaknesses?
Its strongest dimension is Substance & Value (58th on the board, 82.4). Its weakest is Length Discipline (77th on the board, 66.5). The full nine-metric breakdown is on this page.
Does Inkling write some formats better than others?
Yes. Its best of our 9 scripts was the founder announcement / personal origin story (84.4) and its weakest was the personal technical walkthrough / agentic workflow case study (71.8).
How consistent is Inkling between runs?
We run every script 5 times. Inkling's overall score varies by about ±4.5 points between runs, versus a board median of ±3.8. That is typical consistency for this board. The full per-metric variability table is on this page.
How was Inkling evaluated?
Via Together AI (serverless) using the exact model/route id thinkingmachines/Inkling, run on 2026-07-29. 5 runs per script, provider-default sampling, no fine-tuning; every draft scored blind with a fixed rubric by a three-family judge panel. Full details in the 'How we ran it' section and the methodology.
Is Inkling open source?
Yes, it is an open-weights model. Among open-weights models it ranks 12th of 47 for writing.