NATIONS ROYALE BENCHMARK · REASONING MEDIUM

Claude Sonnet 5.5

Anthropic · released 2026-09-28 · route claude-sonnet-5-5 · Model profile · Chess results · Sudoku results · Mines results · Nonograms results · FreeCell results · Starjar results · Wildhaven results · Hustle Sim results · Elemental Armory results · Micro Mouse results · Ravelin results

Results

Score
8395%: 65 to 100
Wars won
10 of 12
Won / lost / timed out
10 / 2 / 0
Share of the world held at the end, on average
61.8%
Decisions per war
70.6
Orders sent per decision
3.60
Orders the game refused
82 of 3,053 (2.7%)
Refused orders per decision
0.10
Decisions with no readable orders
0.0%
Retries for a reply with no orders tag
0
Model minutes per war
7.1
Tokens per decision (in / out)
3,889 / 601
Reasoning tokens per decision
494
Game build
0.91.0+e426bd2
API-equivalent cost per war
$1.32
Seconds per decision
6.1
Replies cut off at the length limit
0.0%
Aborted on our side (not counted)
52

By tier

TierWarsScoreWon / lost / timed outHeldMedian winReference player
Easy31003 / 0 / 075%3:54100
Medium31003 / 0 / 073%5:2967
Hard31003 / 0 / 075%5:5533
Extreme3331 / 2 / 024%9:470

Against the game's players, war by war

Its war score minus each player's on the same wars, in points out of 100. Better and worse count the wars where it scored higher or lower.

PlayerWarsGapBetter / worse
Reference player12+33.34 / 0
Reference on the model's clock12+51.87 / 0
Sharp persona12+25.03 / 0
Median persona12+42.56 / 0
Tourist persona12+66.28 / 0

Every war