0

LiveBench - Language

Frontier

Language sub-leaderboard of LiveBench: typo correction, plot unscrambling, NYT Connections.

Open
Publisher
Abacus.AI
Published
May 2026
Canonical
livebench.ai

Cite

Notes

Only stored in your browser.

Attribution

Leaderboard scores
LiveBench
Attribution policy →

Top score 85.6% by GPT-5.5 - 137 models reporting (59 frontier)

Score history

100
0%25%50%75%100%Nov 22Aug 23May 24Feb 25Nov 25GPT-3.5 TurboGPT-4GPT-4 TurboClaude Opus 3o1 PreviewGPT-5.1GPT-5.5

Top models

137
LiveBench - LanguageBar chart with 21 bars. Highest value: GPT-5.5 at 85.6.
21 models

FAQ

What is LiveBench - Language?
Language sub-leaderboard of LiveBench: typo correction, plot unscrambling, NYT Connections.
What is the current top score on LiveBench - Language?
The top reported score is 85.6% by GPT-5.5, across 137 models reporting (59 from frontier labs).