Towards AITowards AIToneBench

Thinking Levels · Anthropic

Claude Opus 5.5: Writing Quality by Reasoning Effort

Last updated , from board data generated at 17:52 UTC.

This page compares Claude Opus 5.5 at 5 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, Claude Opus 5.5 ranks #1 of 87 models through its best setting. The other ranks on this page are configuration ranks: one setting's place among all 151 configurations.

What More Effort Changes for Claude Opus 5.5

Claude Opus 5.5 gains a lot from more effort, and nearly all of it comes in one step. Low, medium and high sit in one overlapping cluster near the top of the board, then xhigh jumps 233 Elo past high and lands at #1 by Writing Elo. From low to max, the climb adds up to 274 Elo, enough for max to beat low in about 83% of script comparisons. The scores show where the jump comes from: at xhigh, the scripts land much closer to their target length, and every other score inches up with it.

Recommended Setting

I'd run it at xhigh. It's #1 by Writing Elo, at $1.00 and about 6.7 minutes a script. Max sits right behind it at #2, 5 Elo lower on the point estimate, and the two ranges overlap, though overlap alone doesn't test their difference. The bill separates them: max costs 2.6 times as much and keeps you waiting 17.5 minutes per script, and I don't see what that buys.

If xhigh is too slow or too pricey for your volume, drop to Claude Opus 5.5 (high). It's still #8 on the board for $0.328, and xhigh costs 3.1 times as much. Below high, the ranges overlap again. I'd use low for rough drafts too.

Every Setting on the Chart

Each point is one effort setting of Claude Opus 5.5, joined from the lowest setting to the highest. Higher and further left is better: a higher score for less money or less time.

How to Use the Chart
  • A hollow marker is the provider default: the level the provider applies when the request sets no effort level.
  • The tabs choose the score on the vertical axis, X axis switches between cost and time, and Table lists the same numbers as rows.
  • Select a point for its numbers and a link to its scorecard.
Score on the vertical axis
View
X axis

On a small screen, Table lists the same numbers as rows.

All Effort Settings, Step by Step

Score colors:82 and up65 to 81.9under 65
SettingWriting EloStepWriting ScoreCost / ScriptTime / ScriptConfiguration Rank
low2425 ±60-84.3$0.2431.5 min#10
medium2452 ±55+2784.5$0.3101.9 min#9
high2471 ±50+1984.6$0.3282.1 min#8
xhigh2704 ±51+23385.8$1.006.7 min#1
max2699 ±58-585.9$2.5817.5 min#2

Disclosure: Claude Opus 5.5 is also one of the four judges, so it scores the drafts of its own model line on this page. The other three judges, two LLM judges from other labs and Jev, score every one of those drafts too, so no draft is scored by its own family alone. The methodology describes the judge panel, and the judge analysis measures how much each judge favors its own lab.

Other Anthropic Models by Effort

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 previous explicit setting. 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 Anthropic family page lists every Anthropic model side by side.