math benchmark
LiveBench
LiveBench is a challenging, contamination-limited LLM benchmark that addresses test set contamination by releasing new questions monthly based on recently-released datasets, arXiv papers, news articles, and IMDb movie synopses. It comprises tasks across math, coding, reasoning, language, instruction following, and data analysis with verifiable, objective ground-truth answers.
Updated Aug 7, 2026
Published models38
Registry coverage38
MetricScore
EvidenceB
LiveBench leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
Score distribution
Top published rows on the benchmark's original scale.
LiveBench highlights
The top published results on this benchmark's own scale.
What is LiveBench?
Definition and scoring fields from the benchmark registry.
LiveBench is a challenging, contamination-limited LLM benchmark that addresses test set contamination by releasing new questions monthly based on recently-released datasets, arXiv papers, news articles, and IMDb movie synopses. It comprises tasks across math, coding, reasoning, language, instruction following, and data analysis with verifiable, objective ground-truth answers.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- LiveBench
- Modality
- text
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- livebench|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about LiveBench.
Which model scores highest on LiveBench?
o3-mini is currently ranked first with 84.6%.
What does LiveBench measure?
LiveBench is a challenging, contamination-limited LLM benchmark that addresses test set contamination by releasing new questions monthly based on recently-released datasets, arXiv papers, news articles, and IMDb movie synopses. It comprises tasks across math, coding, reasoning, language, instruction following, and data analysis with verifiable, objective ground-truth answers.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
38 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.