Skip to main content
LIVE RADAR
OpenAI
Autonomous Risk Evaluation System Hallucination Drift
#ALP-2026
(Verified)
Anthropic
Biased Candidate Scoring Algorithm Suppression
#ALP-2026
(Verified)
Google AI
RAG Vector Store Corporate Document Leakage
#ALP-2026
(Verified)
DeepSeek
Credit Underwriting Blackbox Decision Drift
#ALP-2026
(Verified)
OpenAI
Autonomous Risk Evaluation System Hallucination Drift
#ALP-2026
(Verified)
Anthropic
Biased Candidate Scoring Algorithm Suppression
#ALP-2026
(Verified)
Google AI
RAG Vector Store Corporate Document Leakage
#ALP-2026
(Verified)
DeepSeek
Credit Underwriting Blackbox Decision Drift
#ALP-2026
(Verified)
Skip to main content
ALPAR
AI
Justice Board
Leaderboard
Dilemmas
Transparency
Products
EN
Log a Record
Sign in
Feed
Incidents
Ranks
Dilemmas
Claude 4 Sonnet — Score History
Time-series view of K-BENCHMARK category scores for this model.
Ethics & Safety
Score: 8400
Samples: 100
Range: 7600–9000
Hallucination & Factuality
Score: 8100
Samples: 100
Range: 7200–8700
Turkish Competence
Score: 8000
Samples: 100
Range: 7100–8700
EU AI Act Reasoning
Score: 8200
Samples: 100
Range: 7300–8800
Math & Reasoning
Score: 8800
Samples: 100
Range: 8000–9300
Instruction Following
Score: 8800
Samples: 100
Range: 8000–9300
Robustness & Adversarial
Score: 8000
Samples: 100
Range: 7100–8700
Long-Context Retrieval
Score: 8400
Samples: 100
Range: 7600–9000