{"record": {"aliases": ["SEC-Bench Pro"], "artifacts": [{"kind": "repo", "url": "https://github.com/sec-bench/sec-bench"}], "audits": [], "categories": ["security"], "description": {"en": "Security engineering benchmark; evaluates both vulnerability detection and secure coding practices."}, "documents": [{"document_type": "technical_report", "id": "model_reports:deepseek_v4_1_flash_technical_report", "model_name": "DeepSeek-V4.1-Flash", "organization": "DeepSeek", "published": "2026-09-10", "retrieved_at": "2026-09-11", "source": "model_reports", "source_id": "deepseek_v4_1_flash_technical_report", "source_url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/DeepSeek_V41_Tech_Report.pdf", "title": "DeepSeek-V4.1-Flash"}], "evidence_summary": {"document_count": 1, "model_count": 1, "model_count_basis": "source_model_id"}, "key": "model-reports:sec_bench_pro", "modality": null, "name": "SEC-Bench Pro", "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://github.com/sec-bench/sec-bench"}, "publisher": null, "released": null, "released_reference": null, "schema_version": 1, "sizes": [], "slug": "sec_bench_pro", "source": "model_reports", "source_benchmark_id": "sec_bench_pro"}, "schema_version": 1, "scores_by_source": {"model_reports": {"rows": [{"benchmark_id": "sec_bench_pro", "comparable_group": "[\"model-reports:sec_bench_pro\",\"sec_bench_pro\",\"think max, Pass@1\"]", "date_precision": "document_publication", "document_id": "model_reports:deepseek_v4_1_flash_technical_report", "instrument": "sec_bench_pro", "key": "model-reports:sec_bench_pro", "measured_by": "DeepSeek", "measurement_kind": "reported_self_score", "model": "DeepSeek-V4.1-Flash", "model_id": "[\"DeepSeek\",\"DeepSeek-V4.1-Flash\"]", "model_name": "DeepSeek-V4.1-Flash", "obs_id": "curated\u0000sec_bench_pro\u0000deepseek_v4_1_flash_technical_report\u0000sec_bench_pro\u0000think max, Pass@1\u0000DeepSeek-V4.1-Flash", "observation_id": "curated\u0000sec_bench_pro\u0000deepseek_v4_1_flash_technical_report\u0000sec_bench_pro\u0000think max, Pass@1\u0000DeepSeek-V4.1-Flash", "organization": "DeepSeek", "protocol": "think max, Pass@1", "rank_in_source_response": null, "raw_value": "62.8", "read_from": "pdf_text", "reported_at": "2026-09-10", "reported_by": "self_reported", "reported_date": "2026-09-10", "series_id": "model_reports:sec_bench_pro:default", "source": "model_reports", "source_document_type": "technical_report", "source_id": "deepseek_v4_1_flash_technical_report", "source_title": "DeepSeek-V4.1-Flash", "source_url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/DeepSeek_V41_Tech_Report.pdf", "value": 62.8, "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\u0000sec_bench_pro\u0000deepseek_v4_1_flash_technical_report\u0000sec_bench_pro\u0000think max, Pass@1\u0000DeepSeek-V4.1-Flash", "reported_at": "2026-09-10", "source_url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/DeepSeek_V41_Tech_Report.pdf"}, "first_score_report": {"date_precision": "document_publication", "obs_id": "curated\u0000sec_bench_pro\u0000deepseek_v4_1_flash_technical_report\u0000sec_bench_pro\u0000think max, Pass@1\u0000DeepSeek-V4.1-Flash", "reported_at": "2026-09-10", "source_url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/DeepSeek_V41_Tech_Report.pdf"}, "key": "model-reports:sec_bench_pro", "max_score_contradicted": false, "metric": "accuracy", "observation_count": 1, "observed_max": 62.8, "score_summary": {"display_max": 62.8, "display_multiplier": 1, "model_count": 1, "model_count_basis": "source_model_id", "numeric_count": 1, "raw_max": 62.8, "raw_min": 62.8, "source_reference": {"obs_id": "curated\u0000sec_bench_pro\u0000deepseek_v4_1_flash_technical_report\u0000sec_bench_pro\u0000think max, Pass@1\u0000DeepSeek-V4.1-Flash", "observation_id": "curated\u0000sec_bench_pro\u0000deepseek_v4_1_flash_technical_report\u0000sec_bench_pro\u0000think max, Pass@1\u0000DeepSeek-V4.1-Flash", "reported_at": "2026-09-10", "reported_date": "2026-09-10", "source_id": "deepseek_v4_1_flash_technical_report", "source_url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/DeepSeek_V41_Tech_Report.pdf"}, "unit": "percent"}, "series_id": "model_reports:sec_bench_pro:default", "unit": "percent"}}}, "siblings": []}