{"record": {"aliases": ["WebsiteBench", "WebsiteBench: Can AI Agents Rebuild Websites through Browser Exploration?"], "artifacts": [{"kind": "website", "url": "https://website-bench.com/"}], "audits": [], "categories": ["coding_agent"], "description": {"en": "Agents rebuild production websites after exploring them in a browser. The leaderboard reports hidden-test pass rates across 70 tasks. Each task has 302 to 2,016 cases; a case passes only when its functional checks pass and every visual checkpoint reaches normalized SSIM 0.99. Overall pools passed and total tests across tasks rather than averaging task rewards. The table does not state model run dates, harnesses or execution budgets."}, "documents": [{"document_type": "benchmark_leaderboard", "id": "model_reports:websitebench_leaderboard_2026_10_17", "model_name": null, "organization": "TIGER-AI-Lab", "published": "2026-10-17", "retrieved_at": "2026-11-12", "source": "model_reports", "source_id": "websitebench_leaderboard_2026_10_17", "source_url": "https://website-bench.com/#results", "title": "WebsiteBench official leaderboard (2026-10-17 cohort)"}], "evidence_summary": {"document_count": 1, "model_count": 8, "model_count_basis": "source_model_id"}, "key": "model-reports:websitebench", "modality": null, "name": "WebsiteBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"registry": "data/model_cards.yml", "score_archive": "data/benchmark_scores.yml", "source_url": "https://website-bench.com/"}, "publisher": null, "released": null, "released_reference": null, "schema_version": 1, "sizes": [], "slug": "websitebench", "source": "model_reports", "source_benchmark_id": "websitebench"}, "schema_version": 1, "scores_by_source": {"model_reports": {"rows": [{"benchmark_id": "websitebench", "comparable_group": "[\"model-reports:websitebench\",\"websitebench_70_20261017\",\"websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5.5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\"]", "date_precision": "document_publication", "document_id": "model_reports:websitebench_leaderboard_2026_10_17", "instrument": "websitebench_70_20261017", "key": "model-reports:websitebench", "measured_by": "TIGER-AI-Lab", "measurement_kind": "benchmark_publisher_run", "model": "Claude Opus 5.5", "model_id": "[\"Anthropic\",\"Claude Opus 5.5\"]", "model_name": "Claude Opus 5.5", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5.5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Claude Opus 5.5", "observation_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5.5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Claude Opus 5.5", "organization": "Anthropic", "protocol": "websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5.5; exact context, reasoning, harness and run dates unavailable in current publisher metadata", "rank_in_source_response": null, "raw_value": "26.2", "read_from": "html_text", "reported_at": "2026-10-17", "reported_by": "third_party", "reported_date": "2026-10-17", "series_id": "model_reports:websitebench:default", "source": "model_reports", "source_document_type": "benchmark_leaderboard", "source_id": "websitebench_leaderboard_2026_10_17", "source_title": "WebsiteBench official leaderboard (2026-10-17 cohort)", "source_url": "https://website-bench.com/#results", "value": 26.2, "value_kind": "number"}, {"benchmark_id": "websitebench", "comparable_group": "[\"model-reports:websitebench\",\"websitebench_70_20261017\",\"websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\"]", "date_precision": "document_publication", "document_id": "model_reports:websitebench_leaderboard_2026_10_17", "instrument": "websitebench_70_20261017", "key": "model-reports:websitebench", "measured_by": "TIGER-AI-Lab", "measurement_kind": "benchmark_publisher_run", "model": "Claude Opus 5", "model_id": "[\"Anthropic\",\"Claude Opus 5\"]", "model_name": "Claude Opus 5", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Claude Opus 5", "observation_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Claude Opus 5", "organization": "Anthropic", "protocol": "websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5; exact context, reasoning, harness and run dates unavailable in current publisher metadata", "rank_in_source_response": null, "raw_value": "17.1", "read_from": "html_text", "reported_at": "2026-10-17", "reported_by": "third_party", "reported_date": "2026-10-17", "series_id": "model_reports:websitebench:default", "source": "model_reports", "source_document_type": "benchmark_leaderboard", "source_id": "websitebench_leaderboard_2026_10_17", "source_title": "WebsiteBench official leaderboard (2026-10-17 cohort)", "source_url": "https://website-bench.com/#results", "value": 17.1, "value_kind": "number"}, {"benchmark_id": "websitebench", "comparable_group": "[\"model-reports:websitebench\",\"websitebench_70_20261017\",\"websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: DeepSeek V4.1 Flash; exact context, reasoning, harness and run dates unavailable in current publisher metadata\"]", "date_precision": "document_publication", "document_id": "model_reports:websitebench_leaderboard_2026_10_17", "instrument": "websitebench_70_20261017", "key": "model-reports:websitebench", "measured_by": "TIGER-AI-Lab", "measurement_kind": "benchmark_publisher_run", "model": "DeepSeek-V4.1-Flash", "model_id": "[\"DeepSeek\",\"DeepSeek-V4.1-Flash\"]", "model_name": "DeepSeek-V4.1-Flash", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: DeepSeek V4.1 Flash; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000DeepSeek-V4.1-Flash", "observation_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: DeepSeek V4.1 Flash; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000DeepSeek-V4.1-Flash", "organization": "DeepSeek", "protocol": "websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: DeepSeek V4.1 Flash; exact context, reasoning, harness and run dates unavailable in current publisher metadata", "rank_in_source_response": null, "raw_value": "13.5", "read_from": "html_text", "reported_at": "2026-10-17", "reported_by": "third_party", "reported_date": "2026-10-17", "series_id": "model_reports:websitebench:default", "source": "model_reports", "source_document_type": "benchmark_leaderboard", "source_id": "websitebench_leaderboard_2026_10_17", "source_title": "WebsiteBench official leaderboard (2026-10-17 cohort)", "source_url": "https://website-bench.com/#results", "value": 13.5, "value_kind": "number"}, {"benchmark_id": "websitebench", "comparable_group": "[\"model-reports:websitebench\",\"websitebench_70_20261017\",\"websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: GLM 5.3 Flash; exact context, reasoning, harness and run dates unavailable in current publisher metadata\"]", "date_precision": "document_publication", "document_id": "model_reports:websitebench_leaderboard_2026_10_17", "instrument": "websitebench_70_20261017", "key": "model-reports:websitebench", "measured_by": "TIGER-AI-Lab", "measurement_kind": "benchmark_publisher_run", "model": "GLM-5.3-Flash", "model_id": "[\"Z.ai\",\"GLM-5.3-Flash\"]", "model_name": "GLM-5.3-Flash", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: GLM 5.3 Flash; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000GLM-5.3-Flash", "observation_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: GLM 5.3 Flash; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000GLM-5.3-Flash", "organization": "Z.ai", "protocol": "websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: GLM 5.3 Flash; exact context, reasoning, harness and run dates unavailable in current publisher metadata", "rank_in_source_response": null, "raw_value": "14.7", "read_from": "html_text", "reported_at": "2026-10-17", "reported_by": "third_party", "reported_date": "2026-10-17", "series_id": "model_reports:websitebench:default", "source": "model_reports", "source_document_type": "benchmark_leaderboard", "source_id": "websitebench_leaderboard_2026_10_17", "source_title": "WebsiteBench official leaderboard (2026-10-17 cohort)", "source_url": "https://website-bench.com/#results", "value": 14.7, "value_kind": "number"}, {"benchmark_id": "websitebench", "comparable_group": "[\"model-reports:websitebench\",\"websitebench_70_20261017\",\"websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: GPT 6.1 Sol; exact context, reasoning, harness and run dates unavailable in current publisher metadata\"]", "date_precision": "document_publication", "document_id": "model_reports:websitebench_leaderboard_2026_10_17", "instrument": "websitebench_70_20261017", "key": "model-reports:websitebench", "measured_by": "TIGER-AI-Lab", "measurement_kind": "benchmark_publisher_run", "model": "GPT-6.1 Sol", "model_id": "[\"OpenAI\",\"GPT-6.1 Sol\"]", "model_name": "GPT-6.1 Sol", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: GPT 6.1 Sol; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000GPT-6.1 Sol", "observation_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: GPT 6.1 Sol; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000GPT-6.1 Sol", "organization": "OpenAI", "protocol": "websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: GPT 6.1 Sol; exact context, reasoning, harness and run dates unavailable in current publisher metadata", "rank_in_source_response": null, "raw_value": "20.2", "read_from": "html_text", "reported_at": "2026-10-17", "reported_by": "third_party", "reported_date": "2026-10-17", "series_id": "model_reports:websitebench:default", "source": "model_reports", "source_document_type": "benchmark_leaderboard", "source_id": "websitebench_leaderboard_2026_10_17", "source_title": "WebsiteBench official leaderboard (2026-10-17 cohort)", "source_url": "https://website-bench.com/#results", "value": 20.2, "value_kind": "number"}, {"benchmark_id": "websitebench", "comparable_group": "[\"model-reports:websitebench\",\"websitebench_70_20261017\",\"websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Grok 4.6; exact context, reasoning, harness and run dates unavailable in current publisher metadata\"]", "date_precision": "document_publication", "document_id": "model_reports:websitebench_leaderboard_2026_10_17", "instrument": "websitebench_70_20261017", "key": "model-reports:websitebench", "measured_by": "TIGER-AI-Lab", "measurement_kind": "benchmark_publisher_run", "model": "Grok 4.6", "model_id": "[\"xAI\",\"Grok 4.6\"]", "model_name": "Grok 4.6", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Grok 4.6; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Grok 4.6", "observation_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Grok 4.6; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Grok 4.6", "organization": "xAI", "protocol": "websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Grok 4.6; exact context, reasoning, harness and run dates unavailable in current publisher metadata", "rank_in_source_response": null, "raw_value": "12.1", "read_from": "html_text", "reported_at": "2026-10-17", "reported_by": "third_party", "reported_date": "2026-10-17", "series_id": "model_reports:websitebench:default", "source": "model_reports", "source_document_type": "benchmark_leaderboard", "source_id": "websitebench_leaderboard_2026_10_17", "source_title": "WebsiteBench official leaderboard (2026-10-17 cohort)", "source_url": "https://website-bench.com/#results", "value": 12.1, "value_kind": "number"}, {"benchmark_id": "websitebench", "comparable_group": "[\"model-reports:websitebench\",\"websitebench_70_20261017\",\"websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Kimi K3; exact context, reasoning, harness and run dates unavailable in current publisher metadata\"]", "date_precision": "document_publication", "document_id": "model_reports:websitebench_leaderboard_2026_10_17", "instrument": "websitebench_70_20261017", "key": "model-reports:websitebench", "measured_by": "TIGER-AI-Lab", "measurement_kind": "benchmark_publisher_run", "model": "Kimi K3", "model_id": "[\"Moonshot AI\",\"Kimi K3\"]", "model_name": "Kimi K3", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Kimi K3; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Kimi K3", "observation_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Kimi K3; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Kimi K3", "organization": "Moonshot AI", "protocol": "websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Kimi K3; exact context, reasoning, harness and run dates unavailable in current publisher metadata", "rank_in_source_response": null, "raw_value": "10.6", "read_from": "html_text", "reported_at": "2026-10-17", "reported_by": "third_party", "reported_date": "2026-10-17", "series_id": "model_reports:websitebench:default", "source": "model_reports", "source_document_type": "benchmark_leaderboard", "source_id": "websitebench_leaderboard_2026_10_17", "source_title": "WebsiteBench official leaderboard (2026-10-17 cohort)", "source_url": "https://website-bench.com/#results", "value": 10.6, "value_kind": "number"}, {"benchmark_id": "websitebench", "comparable_group": "[\"model-reports:websitebench\",\"websitebench_70_20261017\",\"websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Qwen 3.8 Max; exact context, reasoning, harness and run dates unavailable in current publisher metadata\"]", "date_precision": "document_publication", "document_id": "model_reports:websitebench_leaderboard_2026_10_17", "instrument": "websitebench_70_20261017", "key": "model-reports:websitebench", "measured_by": "TIGER-AI-Lab", "measurement_kind": "benchmark_publisher_run", "model": "Qwen 3.8 Max", "model_id": "[\"Qwen\",\"Qwen 3.8 Max\"]", "model_name": "Qwen 3.8 Max", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Qwen 3.8 Max; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Qwen 3.8 Max", "observation_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Qwen 3.8 Max; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Qwen 3.8 Max", "organization": "Qwen", "protocol": "websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Qwen 3.8 Max; exact context, reasoning, harness and run dates unavailable in current publisher metadata", "rank_in_source_response": null, "raw_value": "7.3", "read_from": "html_text", "reported_at": "2026-10-17", "reported_by": "third_party", "reported_date": "2026-10-17", "series_id": "model_reports:websitebench:default", "source": "model_reports", "source_document_type": "benchmark_leaderboard", "source_id": "websitebench_leaderboard_2026_10_17", "source_title": "WebsiteBench official leaderboard (2026-10-17 cohort)", "source_url": "https://website-bench.com/#results", "value": 7.3, "value_kind": "number"}], "series": {"declared_max": null, "direction": "higher_is_better", "direction_basis": "reported_metric", "display_scale": null, "first_score_record": {"date_precision": "document_publication", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5.5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Claude Opus 5.5", "reported_at": "2026-10-17", "source_url": "https://website-bench.com/#results"}, "first_score_report": {"date_precision": "document_publication", "obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5.5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Claude Opus 5.5", "reported_at": "2026-10-17", "source_url": "https://website-bench.com/#results"}, "key": "model-reports:websitebench", "max_score_contradicted": false, "metric": "hidden_test_pass_rate", "observation_count": 8, "observed_max": 26.2, "score_summary": {"display_max": 26.2, "display_multiplier": 1, "model_count": 8, "model_count_basis": "source_model_id", "numeric_count": 8, "raw_max": 26.2, "raw_min": 7.3, "source_reference": {"obs_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5.5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Claude Opus 5.5", "observation_id": "curated\u0000websitebench\u0000websitebench_leaderboard_2026_10_17\u0000websitebench_70_20261017\u0000websitebench-70-20261017 cohort; 70 tasks; test-pooled 100 * sum(passed tests) / sum(total tests); Overall; model configuration: Claude Opus 5.5; exact context, reasoning, harness and run dates unavailable in current publisher metadata\u0000Claude Opus 5.5", "reported_at": "2026-10-17", "reported_date": "2026-10-17", "source_id": "websitebench_leaderboard_2026_10_17", "source_url": "https://website-bench.com/#results"}, "unit": "percent"}, "series_id": "model_reports:websitebench:default", "unit": "percent"}}}, "siblings": []}