The Arena Leaderboard
AI Rankings
BigAIArena’s official leaderboard. Every score is calculated from comparisons with publicly verifiable evidence.
Rankings are calculated from the average of 5 criteria across comparisons played. The higher the score, the fewer Hallucination, Fake Citation, and Garbage failures.
Public Ledger
Leaderboard
| Rank | AI Model | Score | Comparisons | Verified Comparisons | ELO | Status |
|---|---|---|---|---|---|---|
| 1 | Coming Soon | – | – | – | – | – |
How It’s Measured
Ranking Metrics
| Metric | Description |
|---|---|
| Total Score | Average score across comparisons. |
| Comparisons | Total comparisons played. |
| Verified Comparisons | Comparisons scoring Arena Verified (80+) or higher. |
| ELO Rating | Ranking adjusted for the strength of the AIs a given AI has been paired against. |
| Stability Score | Degree of consistency across comparisons. |
| Evidence Score | Quality of citations and evidence. |
Score To Status
Arena Status Levels
| Score | Status |
|---|---|
| 95–100 | Hall Of Truth Elite |
| 90–94 | Hall Of Truth |
| 80–89 | Arena Verified |
| 40–79 | Arena Pass |
| 0–39 | Improvement Required |
Filter By Criterion
Ranking Categories
Overall · Citation · Anti-Hallucination · Anti-Sycophancy · Anti-Garbage · Post-Comparison Audit · Evidence Quality · Stability
Top Tier
Hall Of Truth — Minimum Requirements
Score of 90 or above · No severe Fake Citation · No severe Hallucination · Verifiable citations · At least 10 completed comparisons.
Bottom Tier
Improvement Required — Common Causes
Hallucination · Fake Citation · Fake Web Access · Sycophancy · Garbage Output.
Consistency Over Time
Stability Score
| Level | Description |
|---|---|
| Excellent | Consistent across many comparisons. |
| Good | Minor fluctuation. |
| Average | Inconsistent results. |
| Poor | Large swings between comparisons. |
Refine The View
Ranking Filters
AI Model · Date Range (Last 30 Days / Last 100 Comparisons) · Status · Score Range · Website Category
Coming Later
Future Integration
Direct links to Comparison Results, Arena Cases, Open Verification, BrainCrisis Eco, RealDataset Hub.
Standard
Rankings Are Earned
No Model Is Above Evidence.