Thinking Levels · OpenAI
Last updated , from board data generated at 17:52 UTC.
This page compares gpt-oss 120B at 2 reasoning-effort settings, each tested on the same 10 scripts: the Writing Score (the 0 to 100 consensus of the judges and code), the Writing Elo (a rating from pairwise comparisons of those scores), and the cost and time of each setting. As a model, gpt-oss 120B ranks #75 of 87 models through its best setting. The other ranks on this page are configuration ranks: one setting's place among all 151 configurations.
High reasoning gives gpt-oss 120B a lift on Writing Elo. Asking for high lifts it 34 Elo over the medium setting you get with no flag, but the two ranges overlap, and both settings sit near the bottom of the board, at #138 and #140. What does move is the bill and the wait: 2.0 times the price per script, and 1.6 minutes instead of 0.5.
I'd keep it on medium. Both prices are a fraction of a cent, so money isn't really the question. The extra reasoning mostly buys a script that lands closer to the target length and has better visual cues, while tone and substance stay about where they were. And tone is the real problem with this model: Tone & Voice Match, the metric that weighs the most, stays low at both settings, so you'd be rewriting most of the voice anyway.
It still makes sense if you need open weights you can run yourself, for privacy for example. For scripts, though, GLM-5.3 Flash (max) is open too and sits at #29 for $0.022 a script. That's where I'd start.
Each point is one effort setting of gpt-oss 120B, joined from the lowest setting to the highest. Higher and further left is better: a higher score for less money or less time.
On a small screen, Table lists the same numbers as rows.
| Setting | Writing Elo | Step | Writing Score | Cost / Script | Time / Script | Reasoning Tokens | Configuration Rank |
|---|---|---|---|---|---|---|---|
| medium (provider default) | 507 ±52 | - | 60.1 | $0.0036 | 0.5 min | 567 | #140 |
| high | 541 ±44 | +34 | 61.3 | $0.0073 | 1.6 min | 6,639 | #138 |
The 95% Writing Elo ranges overlap: gpt-oss 120B (medium) runs from 452 to 555, and gpt-oss 120B (high) from 493 to 581.
Writing Elo comes from comparing every pair of models on every script, which puts these settings on the same scale as the leaderboard. Step is the Writing Elo change from the provider default. Cost is one full script at API list prices, and time is the average for one accepted script, retries included. The thinking-levels page covers the other models, and the OpenAI family page lists every OpenAI model side by side.