[ { "model_benchmark_id": 9002, "benchmark_id": "swe-bench-verified", "model_id": "gpt-5-2025-08-07", "score": 0.749, "normalized_score": 0.749, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "Thinking mode enabled (up to 128K tokens) with enhanced reasoning capabilities and iterative problem-solving approach.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "SWE-Bench Verified" }, { "model_benchmark_id": 9004, "benchmark_id": "aider-polyglot", "model_id": "gpt-5-2025-08-07", "score": 0.88, "normalized_score": 0.88, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "Thinking mode enabled (up to 128K tokens) with step-by-step reasoning and multi-language code understanding.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "Aider-Polyglot" }, { "model_benchmark_id": 10027, "benchmark_id": "swe-lancer-(ic-diamond-subset)", "model_id": "gpt-5-2025-08-07", "score": 1.0, "normalized_score": 1.0, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "GPT-5 - IC SWE Diamond Freelance Coding Tasks (earnings-based evaluation).", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "SWE-Lancer (IC-Diamond subset)" }, { "model_benchmark_id": 9020, "benchmark_id": "aime-2025", "model_id": "gpt-5-2025-08-07", "score": 0.946, "normalized_score": 0.946, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 standard with thinking mode enabled (no tools) - competition mathematics.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "AIME 2025" }, { "model_benchmark_id": 9009, "benchmark_id": "mmmu", "model_id": "gpt-5-2025-08-07", "score": 0.842, "normalized_score": 0.842, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode - College-level visual problem-solving with multimodal reasoning.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "MMMU" }, { "model_benchmark_id": 9006, "benchmark_id": "mmlu", "model_id": "gpt-5-2025-08-07", "score": 0.925, "normalized_score": 0.925, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "Standard benchmark across multiple academic subjects with comprehensive knowledge evaluation.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "MMLU" }, { "model_benchmark_id": 9007, "benchmark_id": "humaneval", "model_id": "gpt-5-2025-08-07", "score": 0.934, "normalized_score": 0.934, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "Code generation benchmark with function completion tasks in Python.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "HumanEval" }, { "model_benchmark_id": 9008, "benchmark_id": "math", "model_id": "gpt-5-2025-08-07", "score": 0.847, "normalized_score": 0.847, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "Thinking mode enabled with step-by-step mathematical problem solving and verification.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "MATH" }, { "model_benchmark_id": 9013, "benchmark_id": "healthbench-hard", "model_id": "gpt-5-2025-08-07", "score": 0.016, "normalized_score": 0.016, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "Thinking mode enabled for medical hallucination detection. Measured inaccuracies on challenging healthcare conversations.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "HealthBench Hard" }, { "model_benchmark_id": 9024, "benchmark_id": "frontiermath", "model_id": "gpt-5-2025-08-07", "score": 0.263, "normalized_score": 0.263, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 standard with thinking mode enabled (with python tool only) - FrontierMath Tier 1-3 expert-level mathematics.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "FrontierMath" }, { "model_benchmark_id": 9028, "benchmark_id": "hmmt-2025", "model_id": "gpt-5-2025-08-07", "score": 0.933, "normalized_score": 0.933, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 standard with thinking mode enabled (no tools) - Harvard-MIT Mathematics Tournament.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "HMMT 2025" }, { "model_benchmark_id": 9032, "benchmark_id": "gpqa", "model_id": "gpt-5-2025-08-07", "score": 0.857, "normalized_score": 0.857, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 - Diamond thinking no tools", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "GPQA" }, { "model_benchmark_id": 9037, "benchmark_id": "humanity's-last-exam", "model_id": "gpt-5-2025-08-07", "score": 0.248, "normalized_score": 0.248, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 standard with thinking mode (no tools) - Full set of expert-level questions across subjects.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "Humanity's Last Exam" }, { "model_benchmark_id": 9041, "benchmark_id": "scale-multichallenge", "model_id": "gpt-5-2025-08-07", "score": 0.696, "normalized_score": 0.696, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode enabled - Multi-turn instruction following benchmark.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "Scale MultiChallenge" }, { "model_benchmark_id": 9043, "benchmark_id": "browsecomp", "model_id": "gpt-5-2025-08-07", "score": 0.549, "normalized_score": 0.549, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode enabled - Agentic search & browsing benchmark.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "BrowseComp" }, { "model_benchmark_id": 9045, "benchmark_id": "collie", "model_id": "gpt-5-2025-08-07", "score": 0.99, "normalized_score": 0.99, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode enabled - Instruction-following in freeform writing.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "COLLIE" }, { "model_benchmark_id": 10034, "benchmark_id": "multichallenge-(o3-mini-grader)", "model_id": "gpt-5-2025-08-07", "score": 0.696, "normalized_score": 0.696, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with o3-mini grader - Multi-turn instruction following benchmark with improved grading accuracy.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "MultiChallenge (o3-mini grader)" }, { "model_benchmark_id": 10035, "benchmark_id": "internal-api-instruction-following-(hard)", "model_id": "gpt-5-2025-08-07", "score": 0.64, "normalized_score": 0.64, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "GPT-5 - Internal API instruction following evaluation (hard difficulty).", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "Internal API instruction following (hard)" }, { "model_benchmark_id": 9047, "benchmark_id": "tau2-airline", "model_id": "gpt-5-2025-08-07", "score": 0.626, "normalized_score": 0.626, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "GPT-5 - Function calling benchmark (airline domain).", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "Tau2 airline" }, { "model_benchmark_id": 9049, "benchmark_id": "tau2-retail", "model_id": "gpt-5-2025-08-07", "score": 0.811, "normalized_score": 0.811, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode - Function calling benchmark (retail domain).", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "Tau2 retail" }, { "model_benchmark_id": 9051, "benchmark_id": "tau2-telecom", "model_id": "gpt-5-2025-08-07", "score": 0.967, "normalized_score": 0.967, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode - Function calling benchmark (telecom domain).", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "Tau2 telecom" }, { "model_benchmark_id": 9053, "benchmark_id": "mmmu-pro", "model_id": "gpt-5-2025-08-07", "score": 0.784, "normalized_score": 0.784, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode - Graduate-level visual problem-solving with advanced multimodal reasoning.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "MMMU-Pro" }, { "model_benchmark_id": 9055, "benchmark_id": "videommmu", "model_id": "gpt-5-2025-08-07", "score": 0.846, "normalized_score": 0.846, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode - Video-based multimodal reasoning (max frame 256).", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "VideoMMMU" }, { "model_benchmark_id": 9057, "benchmark_id": "charxiv-r", "model_id": "gpt-5-2025-08-07", "score": 0.811, "normalized_score": 0.811, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode - Scientific figure reasoning and interpretation.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "CharXiv-R" }, { "model_benchmark_id": 9059, "benchmark_id": "erqa", "model_id": "gpt-5-2025-08-07", "score": 0.657, "normalized_score": 0.657, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/gpt-5/", "verified_by_llmstats": false, "analysis_method": "GPT-5 with thinking mode - Multimodal spatial reasoning.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "ERQA" }, { "model_benchmark_id": 10048, "benchmark_id": "openai-mrcr:-2-needle-128k", "model_id": "gpt-5-2025-08-07", "score": 0.952, "normalized_score": 0.952, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "OpenAI-MRCR 2-needle retrieval at 128k tokens.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "OpenAI-MRCR: 2 needle 128k" }, { "model_benchmark_id": 10049, "benchmark_id": "openai-mrcr:-2-needle-256k", "model_id": "gpt-5-2025-08-07", "score": 0.868, "normalized_score": 0.868, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "OpenAI-MRCR 2-needle retrieval at 256k tokens.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "OpenAI-MRCR: 2 needle 256k" }, { "model_benchmark_id": 10050, "benchmark_id": "graphwalks-bfs-<128k", "model_id": "gpt-5-2025-08-07", "score": 0.783, "normalized_score": 0.783, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "Graphwalks BFS (<128k) long-context reasoning.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "Graphwalks BFS <128k" }, { "model_benchmark_id": 10051, "benchmark_id": "graphwalks-parents-<128k", "model_id": "gpt-5-2025-08-07", "score": 0.733, "normalized_score": 0.733, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "Graphwalks parents (<128k) long-context reasoning.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "Graphwalks parents <128k" }, { "model_benchmark_id": 10052, "benchmark_id": "browsecomp-long-128k", "model_id": "gpt-5-2025-08-07", "score": 0.9, "normalized_score": 0.9, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "BrowseComp long-context 128k variant.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "BrowseComp Long Context 128k" }, { "model_benchmark_id": 10053, "benchmark_id": "browsecomp-long-256k", "model_id": "gpt-5-2025-08-07", "score": 0.888, "normalized_score": 0.888, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "BrowseComp long-context 256k variant.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "BrowseComp Long Context 256k" }, { "model_benchmark_id": 10054, "benchmark_id": "videomme-w-sub.", "model_id": "gpt-5-2025-08-07", "score": 0.867, "normalized_score": 0.867, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "VideoMME (long) with subtitles category.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "VideoMME w sub." }, { "model_benchmark_id": 10069, "benchmark_id": "longfact-concepts", "model_id": "gpt-5-2025-08-07", "score": 0.007, "normalized_score": 0.007, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "Thinking mode enabled for hallucination detection. Measured on open-source prompts for concept-based factual queries.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "LongFact-Concepts" }, { "model_benchmark_id": 10070, "benchmark_id": "longfact-objects", "model_id": "gpt-5-2025-08-07", "score": 0.008, "normalized_score": 0.008, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "Thinking mode enabled for hallucination detection. Measured on open-source prompts for object-based factual queries.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "LongFact-Objects" }, { "model_benchmark_id": 10071, "benchmark_id": "factscore", "model_id": "gpt-5-2025-08-07", "score": 0.01, "normalized_score": 0.01, "is_self_reported": true, "self_reported_source_link": "https://openai.com/index/introducing-gpt-5-for-developers/", "verified_by_llmstats": false, "analysis_method": "Thinking mode enabled for factual accuracy assessment. Measured hallucination rate on open-source prompts.", "verification_provider_id": null, "verification_hardware": null, "verification_date": null, "verification_notes": null, "created_at": "2025-07-24T12:00:00.000000+00:00", "updated_at": "2025-07-24T12:00:00.000000+00:00", "benchmark_name": "FactScore" } ]