LiveBench
by LiveBench OSS · free · Last verified 2026-04-24
LiveBench is a contamination-resistant benchmark that continuously updates with new questions sourced from recent math competitions, research papers, and news. By using only data post-dating model training cutoffs, LiveBench mitigates benchmark saturation and provides more reliable capability assessments of frontier models.
https://livebench.ai ↗C
C—Below Average
Adoption: C+Quality: B+Freshness: ACitations: CEngagement: F
Specifications
- License
- Proprietary
- Pricing
- free
- Capabilities
- Integrations
- Use Cases
- API Available
- No
- Evaluated Models
- claude-4, gpt-5, gemini-2.5-pro, deepseek-v3, llama-4-405b
- Metrics
- global-average, math-score, coding-score, reasoning-score
- Methodology
- Monthly-refreshed question sets across 6 categories. All answers programmatically verifiable without LLM judges. Questions sourced from recent events to prevent training contamination.
- Last Run
- 2026-03-15
- Tags
- benchmark, contamination-resistant, live, math, reasoning, continuously-updated
- Added
- 2026-04-24
- Completeness
- 60%
Index Score
44Adoption
50
Quality
70
Freshness
80
Citations
40
Engagement
0