source
LiveBench
Benchmarks · /benchmarks/livebench-language
Language
LiveBench language slice across spelling repair and structured text puzzles.
Passport
Verified but agingThis is an objective signal, so it is mainly about measurable task performance rather than public taste.
metric
Score (%)
judge
Objective
direction
higher better
group id
livebench_language_2026_04
domain
Chat / text
What it measures vs what it misses
✓ Measures
Text-only language handling on objective, contamination-aware tasks.
✗ Misses
Coding workflow quality. Subjective user taste.
Why this countsIt tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.Comparable-group ruleThis percentile only compares models inside the exact benchmark/version group shown here. It is not a universal score.What it missesIt does not prove deeper reasoning, tool use, or enterprise workflow reliability.
Leaderboard · this benchmark version
#1 · o1 Preview
LB · Dec 10, 2024
70.7%
#2 · Gemini 2.5 Pro
LB · Apr 1, 2025
68.5%
#3 · o1
LB · Mar 4, 2025
65.1%
#4 · GPT-4.5 Preview
LB · Apr 3, 2025
63%
#5 · Claude Sonnet 3.7
LB · Apr 1, 2025
62.1%
#6 · Claude Sonnet 3.5
LB · Feb 6, 2025
55.5%
#7 · Grok 3
LB · Mar 18, 2025
53.2%
#8 · DeepSeek Reasoner
LB · Apr 3, 2025
52.1%
#9 · Claude Opus 3
LB · Feb 6, 2025
51.8%
#10 · GPT-4o
LB · Apr 1, 2025
49.7%
#11 · GPT-4
LB · Dec 10, 2024
49.6%
#12 · o3 mini
LB · Feb 6, 2025
48.9%
#13 · Gemini 2.0 Pro Experimental
LB · Apr 1, 2025
48.5%
#14 · Gemini Experimental
LB · Feb 6, 2025
46.8%
#15 · o1 mini
LB · Feb 6, 2025
45.1%
#16 · GPT-4 Turbo
LB · Dec 10, 2024
44.8%
#17 · Grok Beta
LB · Feb 6, 2025
44.8%
#18 · Gemini 1.5 Pro
LB · Feb 6, 2025
43.6%
#19 · Grok 2
LB · Feb 6, 2025
42.4%
#20 · Gemini 2.0 Flash
LB · Apr 7, 2025
41.9%
#21 · Grok 2 mini
LB · Dec 10, 2024
39.5%
#22 · Claude Haiku 3.5
LB · Feb 6, 2025
39.2%
#23 · Claude Haiku 4.5
LB · Feb 6, 2025
39.2%
#24 · Claude Sonnet 3
LB · Dec 10, 2024
38.1%
#25 · Grok 3 mini
LB · Mar 14, 2025
37.8%
#26 · Gemini 2.0 Flash-Lite
LB · Apr 2, 2025
34.5%
#27 · GPT-4o mini
LB · Feb 6, 2025
32.5%
#28 · DeepSeek Chat
LB · Dec 11, 2024
32.1%
#29 · Gemini 1.5 Flash
LB · Dec 12, 2024
30.1%
#30 · Claude Haiku 3
LB · Dec 10, 2024
29.6%
#31 · GPT-3.5 Turbo
LB · Dec 10, 2024
26.4%
#32 · Gemini 1.5 Flash 8B
LB · Apr 3, 2025
21.4%