Thinking Levels · OpenAI
Last updated , from board data generated at 17:52 UTC.
This page compares GPT-5.6 Sol 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-5.6 Sol ranks #11 of 87 models through its best setting. The other ranks on this page are configuration ranks: one setting's place among all 151 configurations.
Ultra effort barely moves GPT-5.6 Sol's Writing Elo rating. The run with no effort flag (low) and the ultra run end up 2 Elo apart, with ranges that overlap almost completely. Ultra still charges you for trying: 3.3 times the price per script, and about 4.1 minutes of waiting instead of 0.9.
Leave the flag off. Ultra barely moves Writing Elo, and the run with no flag costs less and comes back much sooner. If you're picking a model today, though, I'd move to GPT-6 Sol. GPT-6 Sol (max) costs about what this model costs with no flag, $0.126 against $0.113, and ranks well above both settings here, with a clear gap between the ranges.
Each point is one effort setting of GPT-5.6 Sol, 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 | Configuration Rank |
|---|---|---|---|---|---|---|
| low (provider default) | 2132 ±35 | - | 82.3 | $0.113 | 0.9 min | #25 |
| ultra | 2134 ±52 | +2 | 82.3 | $0.367 | 4.1 min | #24 |
The 95% Writing Elo ranges overlap: GPT-5.6 Sol (low) runs from 2095 to 2165, and GPT-5.6 Sol (ultra) from 2082 to 2185.
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.