{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2609.02459", "kind": "paper", "url": "https://arxiv.org/abs/2609.02459"}, {"id": "gh:lmwilki/civ6-mcp", "kind": "repo", "url": "https://github.com/lmwilki/civ6-mcp"}], "authors": ["Austin Tudor David Andrews", "Liam Wilkinson", "Jamie Heagerty", "Harry Coppock", "Jakob Nicolaus Foerster", "Rui Ponte Costa"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2609.02459", "source_url": "https://arxiv.org/abs/2609.02459"}, "categories": ["General AI", "Language & Knowledge", "cs.AI"], "description": {"en": "Evaluates language model agents in long-horizon, tool-mediated Civilization VI gameplay via 76 MCP tools."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2609.02459", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2609.02459", "title": "CivBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["University of Oxford", "Imperial College London"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2609.02459", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2609.02459", "source_url": "https://openalex.org/W7207781962"}, "key": "claire-radar:2609.02459", "modality": null, "name": "CivBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "false", "origin_record_id": "bm_civbench_b296ca93", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "a38cf193664db6a1e5ddaa1810938ebb32a389e44a28b88ffa97f9ebe5354799", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-name-audit-deferred", "reviewed_at": "2026-09-04T00:46:59.007339Z", "source_url": "https://arxiv.org/abs/2609.02459"}, "publisher": null, "released": "2026-09-02", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2609.02459", "source_url": "https://arxiv.org/abs/2609.02459"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2609-02459", "source": "claire_radar", "source_benchmark_id": "2609.02459", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Mixed", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/lmwilki/civ6-mcp", "githubScope": "benchmark_repo", "githubStars": 204, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 0, "hfPaperUrl": "https://huggingface.co/papers/2609.02459", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-10-01", "state": "stale"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "Game-based agent evaluation with open tooling could attract interest from agent and game AI communities.", "score": 65}, "capabilities": [], "confidence": "High", "construction": "Aggregate Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-04T00:46:59.007339Z", "inputHash": "30e530c742cb2ae5fe43ebd4a2a60447cdc9ff9bfe8d0af37083274058e7d60e", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "We present CivBench, an open-source benchmark for evaluating language model agents in long-horizon, tool-mediated environments", "canonicalNameSource": "abstract", "decisionReason": "Code repository is available but the benchmark itself may lack a defined scoring protocol; README emphasizes tool capabilities rather than formal benchmark leaderboard or metrics.", "model": "deepseek-v4-pro", "reviewedAt": "2026-09-04T00:46:59.007339Z", "state": "ai-name-audit-deferred"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates language model agents in long-horizon, tool-mediated Civilization VI gameplay via 76 MCP tools.", "displayEligible": false, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We present CivBench, an open-source benchmark for evaluating language model agents in long-horizon, tool-mediated environments through the Model Context Protocol (MCP)."}, "familyId": "bmf_acb88f2ca1ea", "firstSeenAt": "2026-09-03", "heat": null, "id": "bm_civbench_b296ca93", "indexedAt": "2026-09-03T00:55:28.123857Z", "industrySectors": [], "links": {"code": "https://github.com/lmwilki/civ6-mcp", "data": null, "paper": "https://arxiv.org/abs/2609.02459", "pdf": "https://arxiv.org/pdf/2609.02459", "project": null, "report": "https://arxiv.org/abs/2609.02459"}, "metrics": [], "motivation": "We present CivBench, an open-source benchmark for evaluating language model agents in long-horizon, tool-mediated environments through the Model Context Protocol (MCP).", "name": "CivBench", "oneLine": "Evaluates language model agents in long-horizon, tool-mediated Civilization VI gameplay via 76 MCP tools.", "paperTitle": "CivBench: A Long-Horizon Benchmark for Tool-Mediated Agents in Civilization VI", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2609.02459", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-09-03T00:55:28.123857Z"}, "publications": [], "ranking": {"90d": {"confidence": "Medium", "coverage": 0.7, "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": 0.7105261451429664, "value": 204}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.0, "value": 0}}, "confidence": "Medium", "coverage": 0.7, "rank": 35, "score": 54}, "method": "attention-ranking-v13", "rank": 35, "score": 54}}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-09-02", "source": {"authors": ["Austin Tudor David Andrews", "Liam Wilkinson", "Jamie Heagerty", "Harry Coppock", "Jakob Nicolaus Foerster", "Rui Ponte Costa"], "categories": ["cs.AI"], "id": "2609.02459", "title": "CivBench: A Long-Horizon Benchmark for Tool-Mediated Agents in Civilization VI", "type": "arxiv", "url": "https://arxiv.org/abs/2609.02459"}, "sourceUpdatedAt": "2026-09-02T11:31:39Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "Provides a complex environment for testing long-term strategic reasoning in LLM agents."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
