0

Unscramble

Single-turn transformation where the model unscrambles numbered sentences into the correct order.

Domain
rl-env
License
apache-2.0
Published
Aug 2025

Cite

Notes

Only stored in your browser.

Attribution

Leaderboard scores
prime-hub
Attribution policy →

Top score 45.6% by GPT-4.1 Mini - 1 model reporting (1 frontier)

Top models

1
UnscrambleBar chart with 1 bar. Highest value: GPT-4.1 Mini at 45.6.
1 model

Related tools

1
View all

Implementations, trainers, datasets and scaffolds linked to this eval.

FAQ

What is Unscramble?
Single-turn transformation where the model unscrambles numbered sentences into the correct order.
What is the current top score on Unscramble?
The top reported score is 45.6% by GPT-4.1 Mini, across 1 model reporting (1 from frontier labs).
How can a model improve its Unscramble score?
Tools linked to Unscramble on Sophon include Unscramble RL Env (Prime Intellect) - RL environments, datasets, and scaffolds that target this eval.
What license is Unscramble under?
Unscramble is available under apache-2.0.