{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.03070", "kind": "paper", "url": "https://arxiv.org/abs/2608.03070"}], "authors": ["Jasper Timm", "Lukas Struppek", "Ziwei Xu", "Grace Cheong", "Oscar Mata", "Dan Zhao", "Mick Yang", "Isadora De Andrade", "Xiaojun Jia", "Yiming Li", "Samuel Bauer", "Heather McIntyre", "Adam Gleave", "Edward Yee", "Kellin Pelrine"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.03070", "source_url": "https://arxiv.org/abs/2608.03070"}, "categories": ["Cybersecurity", "Language & Knowledge", "cs.CR"], "description": {"en": "The AI Security Leaderboard evaluates frontier AI model safeguards against the FAR.AI Minimal Standard for Safeguards across severe misuse requests (CBRNE threats and offensive cybersecurity), testing for universal jailbreaks."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.03070", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.03070", "title": "AI Security Leaderboard"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.03070", "modality": null, "name": "AI Security Leaderboard", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_ai-security-leaderboard_0982a1ca", "origin_source": "arxiv", "recognition_confidence": "0.85", "record_sha256": "777de8bdd35271d4510dfc6ac69cfb464eabeb8c356ed742281852f1a877b0a5", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2608.03070"}, "publisher": null, "released": "2026-08-04", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.03070", "source_url": "https://arxiv.org/abs/2608.03070"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-03070", "source": "claire_radar", "source_benchmark_id": "2608.03070", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Cybersecurity"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2608.03070", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "state": "unavailable"}}, "status": "unavailable"}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "69ac3a8722b5e6cde249e6d296b4e83ea963c38877341273c83327b2b789d587", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "The AI Security Leaderboard evaluates frontier AI model safeguards against the FAR.AI Minimal Standard for Safeguards across severe misuse requests (CBRNE threats and offensive cybersecurity), testing for universal jailbreaks.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact coined title identity tied to benchmark evidence", "evaluation protocol evidence"], "snippet": "The AI Security Leaderboard is an independent benchmark that ranks the safeguards of frontier AI models from least to most secure."}, "familyId": "bmf_ee1e051ce1cb", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_ai-security-leaderboard_0982a1ca", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Cybersecurity"], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2608.03070", "paper": "https://arxiv.org/abs/2608.03070", "pdf": "https://arxiv.org/pdf/2608.03070", "project": null, "report": "https://arxiv.org/abs/2608.03070"}, "metrics": [], "motivation": "The AI Security Leaderboard is an independent benchmark that ranks the safeguards of frontier AI models from least to most secure.", "name": "AI Security Leaderboard", "oneLine": "The AI Security Leaderboard evaluates frontier AI model safeguards against the FAR.AI Minimal Standard for Safeguards across severe misuse requests (CBRNE threats and offensive cybersecurity), testing for universal jailbreaks.", "paperTitle": "AI Security Leaderboard: Methodology, Results and Minimal Standard", "primaryDomain": "Cybersecurity", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.03070", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {"90d": {"confidence": "Low", "coverage": 0.0, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": null, "value": null}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "rank": null, "score": null}, "method": "attention-ranking-v13", "rank": null, "score": null}}, "readiness": "Paper only", "recognitionConfidence": 0.85, "relation": "introduces", "releasedAt": "2026-08-04", "source": {"authors": ["Jasper Timm", "Lukas Struppek", "Ziwei Xu", "Grace Cheong", "Oscar Mata", "Dan Zhao", "Mick Yang", "Isadora De Andrade", "Xiaojun Jia", "Yiming Li", "Samuel Bauer", "Heather McIntyre", "Adam Gleave", "Edward Yee", "Kellin Pelrine"], "categories": ["cs.CR", "cs.AI", "cs.CL"], "id": "2608.03070", "title": "AI Security Leaderboard: Methodology, Results and Minimal Standard", "type": "arxiv", "url": "https://arxiv.org/abs/2608.03070"}, "sourceUpdatedAt": "2026-08-05T18:50:24Z", "topics": ["cs.CR"], "venueAttempts": [], "whyItMatters": "Provides a public, rolling comparison of model security that quantifies cost to jailbreak and reveals large gaps, informing deployment decisions for high-risk capabilities."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
