{"record": {"aliases": [], "artifacts": [], "categories": ["reasoning", "spatial_reasoning", "vision"], "description": {"en": "The Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) is a benchmark designed to test general intelligence and abstract reasoning capabilities through visual grid-based transformation tasks. Each task consists of 2-5 demonstration pairs showing input grids transformed into output grids according to underlying rules, with test-takers required to infer these rules and apply them to novel test inputs. The benchmark uses colored grids (up to 30x30) with 10 discrete colors/symbols, designed to measure human-like general fluid intelligence and skill-acquisition efficiency with minimal prior knowledge."}, "documents": [{"document_type": "registry_page", "id": "llm_stats:https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "source": "llm_stats", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "title": "ARC-AGI"}], "evidence_summary": {"document_count": 1, "model_count": 8, "model_count_basis": "source_model_id"}, "key": "llm-stats:arc-agi", "modality": "image", "name": "ARC-AGI", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"crawl_bundle": "llm_stats_2026-08-17", "crawled_at": "2026-08-17T18:49:35.489925+00:00", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500"}, "publisher": null, "released": "2019-11-05", "released_reference": {"basis": "paper_first_version", "note": "First version of the introducing paper: On the Measure of Intelligence. The source-record description identifies this benchmark or its original partition; no dates or scores are joined by name at runtime.", "source_key": "llm-stats:arc-agi", "source_url": "https://arxiv.org/abs/1911.01547"}, "schema_version": 1, "sizes": [], "slug": "llm-stats-arc-agi", "source": "llm_stats", "source_benchmark_id": "arc-agi"}, "schema_version": 1, "scores_by_source": {"llm_stats": {"rows": [{"comparable_group": null, "crawled_at": "2026-08-17T18:49:35.489925+00:00", "date_precision": "model_announcement", "document_id": "llm_stats:https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "key": "llm-stats:arc-agi", "measured_by": null, "model_id": "gpt-5.5", "model_name": "GPT-5.5", "obs_id": "llm_stats:arc-agi:gpt-5.5", "organization": "OpenAI", "rank_in_source_response": 1, "raw_value": "0.95", "reported_by": "self_reported", "reported_date": "2026-04-23", "series_id": "llm_stats:arc-agi:default", "source": "llm_stats", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "value": 0.95, "value_kind": "number"}, {"comparable_group": null, "crawled_at": "2026-08-17T18:49:35.489925+00:00", "date_precision": "model_announcement", "document_id": "llm_stats:https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "key": "llm-stats:arc-agi", "measured_by": null, "model_id": "gpt-5.4", "model_name": "GPT-5.4", "obs_id": "llm_stats:arc-agi:gpt-5.4", "organization": "OpenAI", "rank_in_source_response": 2, "raw_value": "0.937", "reported_by": "self_reported", "reported_date": "2026-03-05", "series_id": "llm_stats:arc-agi:default", "source": "llm_stats", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "value": 0.937, "value_kind": "number"}, {"comparable_group": null, "crawled_at": "2026-08-17T18:49:35.489925+00:00", "date_precision": "model_announcement", "document_id": "llm_stats:https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "key": "llm-stats:arc-agi", "measured_by": null, "model_id": "gpt-5.2-pro-2025-12-11", "model_name": "GPT-5.2 Pro", "obs_id": "llm_stats:arc-agi:gpt-5.2-pro-2025-12-11", "organization": "OpenAI", "rank_in_source_response": 3, "raw_value": "0.905", "reported_by": "self_reported", "reported_date": "2025-12-11", "series_id": "llm_stats:arc-agi:default", "source": "llm_stats", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "value": 0.905, "value_kind": "number"}, {"comparable_group": null, "crawled_at": "2026-08-17T18:49:35.489925+00:00", "date_precision": "model_announcement", "document_id": "llm_stats:https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "key": "llm-stats:arc-agi", "measured_by": null, "model_id": "o3-2025-04-16", "model_name": "o3", "obs_id": "llm_stats:arc-agi:o3-2025-04-16", "organization": "OpenAI", "rank_in_source_response": 4, "raw_value": "0.88", "reported_by": "self_reported", "reported_date": "2025-04-16", "series_id": "llm_stats:arc-agi:default", "source": "llm_stats", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "value": 0.88, "value_kind": "number"}, {"comparable_group": null, "crawled_at": "2026-08-17T18:49:35.489925+00:00", "date_precision": "model_announcement", "document_id": "llm_stats:https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "key": "llm-stats:arc-agi", "measured_by": null, "model_id": "gpt-5.2-2025-12-11", "model_name": "GPT-5.2", "obs_id": "llm_stats:arc-agi:gpt-5.2-2025-12-11", "organization": "OpenAI", "rank_in_source_response": 5, "raw_value": "0.862", "reported_by": "self_reported", "reported_date": "2025-12-11", "series_id": "llm_stats:arc-agi:default", "source": "llm_stats", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "value": 0.862, "value_kind": "number"}, {"comparable_group": null, "crawled_at": "2026-08-17T18:49:35.489925+00:00", "date_precision": "model_announcement", "document_id": "llm_stats:https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "key": "llm-stats:arc-agi", "measured_by": null, "model_id": "inkling-small", "model_name": "Inkling-Small", "obs_id": "llm_stats:arc-agi:inkling-small", "organization": "Thinking Machines Lab", "rank_in_source_response": 6, "raw_value": "0.84", "reported_by": "self_reported", "reported_date": "2026-07-30", "series_id": "llm_stats:arc-agi:default", "source": "llm_stats", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "value": 0.84, "value_kind": "number"}, {"comparable_group": null, "crawled_at": "2026-08-17T18:49:35.489925+00:00", "date_precision": "model_announcement", "document_id": "llm_stats:https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "key": "llm-stats:arc-agi", "measured_by": null, "model_id": "longcat-flash-thinking", "model_name": "LongCat-Flash-Thinking", "obs_id": "llm_stats:arc-agi:longcat-flash-thinking", "organization": "Meituan", "rank_in_source_response": 7, "raw_value": "0.503", "reported_by": "self_reported", "reported_date": "2025-09-22", "series_id": "llm_stats:arc-agi:default", "source": "llm_stats", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "value": 0.503, "value_kind": "number"}, {"comparable_group": null, "crawled_at": "2026-08-17T18:49:35.489925+00:00", "date_precision": "model_announcement", "document_id": "llm_stats:https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "key": "llm-stats:arc-agi", "measured_by": null, "model_id": "qwen3-235b-a22b-instruct-2507", "model_name": "Qwen3-235B-A22B-Instruct-2507", "obs_id": "llm_stats:arc-agi:qwen3-235b-a22b-instruct-2507", "organization": "Qwen", "rank_in_source_response": 8, "raw_value": "0.418", "reported_by": "self_reported", "reported_date": "2025-07-22", "series_id": "llm_stats:arc-agi:default", "source": "llm_stats", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500", "value": 0.418, "value_kind": "number"}], "series": {"bounds": {"basis": "aggregator_declared", "max": 1.0, "min": null}, "declared_max": 1.0, "direction": "higher_is_better", "direction_basis": "source_rank_descending", "display_scale": null, "first_score_record": {"date_precision": "model_announcement", "obs_id": "llm_stats:arc-agi:o3-2025-04-16", "reported_at": "2025-04-16", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500"}, "first_score_report": null, "key": "llm-stats:arc-agi", "max_score_contradicted": false, "metric": null, "observation_count": 8, "observed_max": 0.95, "score_summary": {"display_max": 95.0, "display_multiplier": 100, "model_count": 8, "model_count_basis": "source_model_id", "numeric_count": 8, "raw_max": 0.95, "raw_min": 0.418, "source_reference": {"crawled_at": "2026-08-17T18:49:35.489925+00:00", "obs_id": "llm_stats:arc-agi:gpt-5.5", "reported_date": "2026-04-23", "source_url": "https://api.zeroeval.com/leaderboard/benchmarks/arc-agi?top_n=500"}, "unit": null}, "series_id": "llm_stats:arc-agi:default"}}}, "siblings": []}
