gpt-oss 120B
OpenAI · open weights · writing benchmark
The short version
gpt-oss 120B sits at #113 of 130 on ToneBench, toward the value end of the board, with a writing Elo of 797 and an overall score of 66.8 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 OpenAI it ranks 33rd of 36. GPT-5.6 Sol (ultra) is the family's top writer here, about 1518 Elo ahead, so this config trades some quality for whatever it saves you in cost or speed.
Among open-weights models, it comes in 33rd 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 Length Discipline: it ranks 73rd on that dimension at 67.5, above the board average. Its softer spot is Substance & Value (61.3, 119th), 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 (74.1) and its weakest was the personal technical walkthrough / agentic workflow case study (59.0), a spread of about 15 points. Worth knowing if your writing skews toward one of those formats.
On cost, it runs about $0.001/task. For that money it beats 12 models that cost noticeably more.
If you're tuning reasoning effort, we also tested gpt-oss 120B at other settings. The strongest of those on the board is gpt-oss 120B (high) (Elo 839), so it is worth checking whether more or less thinking moves the needle before you lock in this one.
Skill profile
The shape is the story: the further a corner reaches, the stronger that dimension. Biggest edge over the board: Length Discipline (+0.3); furthest behind: Substance & Value (-17.4).
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, gpt-oss 120B varies by ± 5.2 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 Substance & Value (± 5.2 vs a board median of ± 2.6): two runs of the same brief can land visibly different substance & value scores.
| Metric | This model | Board median | Verdict |
|---|---|---|---|
| Tone & Voice ? | 68.7 ± 3.1 | ± 2.2 | typical |
| Writing Craft ? | 71.4 ± 2.8 | ± 2.0 | typical |
| Substance & Value ? | 61.3 ± 5.2 | ± 2.6 | swingier than most |
| Flow & Emotion ? | 64.1 ± 3.3 | ± 2.6 | typical |
| YouTube Structure ? | 64.7 ± 3.9 | ± 3.6 | typical |
| Hook ? | 72.8 ± 4.0 | ± 3.0 | typical |
| Length Discipline ? | 67.5 ± 9.1 | ± 8.6 | typical |
| Anti-Slop ? | 71.6 ± 2.4 | ± 2.6 | typical |
| Visual Cues ? | 57.2 ± 12.9 | ± 4.7 | swingier than most |
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 66.8 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 15.8 points on its overall. That gap is wider than typical (board median 9.0): DeepSeek V4 Flash rates it noticeably higher than Claude Opus 5. Read the per-metric numbers knowing the consensus sits between two real opinions. How the panel works: methodology.
How we ran gpt-oss 120B
- Accessed viaOpenRouter · model page ↗
- Exact model / route id
openai/gpt-oss-120b(provider default reasoning) - Audited route overrideOpenRouter ·
openai/gpt-oss-120bforautomated_video_workflow(Task-7 exact OpenRouter endpoint pin deepinfra/bf16) - 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 gpt-oss 120B at writing?
On ToneBench it ranks #113 of 130 with a writing Elo of 797 and an overall score of 66.8/100. 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 gpt-oss 120B the best OpenAI model for writing?
Not quite. Within OpenAI it ranks 33rd of 36; GPT-5.6 Sol (ultra) is the family's best writer here.
Is gpt-oss 120B good value for the money?
It costs about $0.001/task. Nothing meaningfully cheaper outscores it, which puts it on the value side of the board.
What are gpt-oss 120B's strengths and weaknesses?
Its strongest dimension is Length Discipline (73rd on the board, 67.5). Its weakest is Substance & Value (119th on the board, 61.3). The full nine-metric breakdown is on this page.
Does gpt-oss 120B write some formats better than others?
Yes. Its best of our 9 scripts was the founder announcement / personal origin story (74.1) and its weakest was the personal technical walkthrough / agentic workflow case study (59.0).
How consistent is gpt-oss 120B between runs?
We run every script 5 times. gpt-oss 120B's overall score varies by about ±5.2 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 gpt-oss 120B evaluated?
Via OpenRouter using the exact model/route id openai/gpt-oss-120b, 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 gpt-oss 120B open source?
Yes, it is an open-weights model. Among open-weights models it ranks 33rd of 47 for writing.