[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"api:\u002Fnav:{}":3,"api:\u002Fbenchmarks\u002Fcl-bench:{\"effort\":\"top\",\"limit\":100,\"page\":1}":104},{"login_code_required":4,"data_date":5,"vendors":6,"categories":39,"topics":55,"model_count":72,"benchmark_count":73,"models":74,"benchmarks":79},false,"2026-10-27",[7,15,23,31],{"name":8,"slug":9,"models":10},"阿里巴巴","alibaba",[11],{"name":12,"slug":13,"short_name":12,"overall_score":14},"Qwen3.8-27B","qwen3-8-27b","47.576875000000000000000000",{"name":16,"slug":17,"models":18},"深度求索","deepseek",[19],{"name":20,"slug":21,"short_name":20,"overall_score":22},"DeepSeek-V4-Pro","deepseek-v4-pro","65.030000000000000000000000",{"name":24,"slug":25,"models":26},"月之暗面","moonshot",[27],{"name":28,"slug":29,"short_name":28,"overall_score":30},"Kimi-K3","kimi-k3","57.474000000000000000000000",{"name":32,"slug":33,"models":34},"智谱","zhipu",[35],{"name":36,"slug":37,"short_name":36,"overall_score":38},"GLM-5.3","glm-5-3","61.067500000000000000000000",[40,44,48,51],{"name":41,"slug":42,"benchmark_count":43},"体验感","experience",2,{"name":45,"slug":46,"benchmark_count":47},"认知能力","cognition",5,{"name":49,"slug":50,"benchmark_count":43},"Agentic 能力","agentic",{"name":52,"slug":53,"benchmark_count":54},"应用领域","domain",1,[56,60,64,68],{"slug":57,"name":58,"question":59},"quant","量化 vs 原精度","量化之后能力掉多少",{"slug":61,"name":62,"question":63},"selfhost","自部署 vs 官方 API","同一份权重，自己部署和调官方 API，分数一致吗",{"slug":65,"name":66,"question":67},"size","同家族尺寸","同一代模型，小尺寸比旗舰差多少",{"slug":69,"name":70,"question":71},"harness","多 Harness","同一个模型换一个 harness，agentic 分数变多少",4,8,[75,76,77,78],{"name":20,"slug":21},{"name":36,"slug":37},{"name":28,"slug":29},{"name":12,"slug":13},[80,83,86,89,92,95,98,101],{"name":81,"slug":82},"AA-LCR","aa-lcr",{"name":84,"slug":85},"AA-Omniscience","aa-omniscience",{"name":87,"slug":88},"AutomationBench","automationbench",{"name":90,"slug":91},"CL-bench","cl-bench",{"name":93,"slug":94},"HLE","hle",{"name":96,"slug":97},"MMLU-Pro","mmlu-pro",{"name":99,"slug":100},"SimpleQA Verified","simpleqa-verified",{"name":102,"slug":103},"SWE-bench Pro","swe-bench-pro",{"benchmark":105,"updated_at":116,"rows":117,"total":54,"categories":150,"tabs":152,"settings":154},{"id":47,"name":90,"slug":91,"license":106,"version":107,"is_active":108,"num_cases":109,"parent_id":107,"score_max":110,"score_min":111,"created_at":112,"has_bundle":108,"is_agentic":4,"sort_order":113,"updated_at":112,"description":107,"metric_name":114,"homepage_url":107,"license_note":107,"scoring_method":107,"composite_weight":110,"higher_is_better":108,"bundle_size_bytes":115},"public",null,true,1899,"1.0000","0.0000","2026-10-28T09:47:38.906402+08:00",0,"Accuracy",728133432,"2026-09-27",[118],{"id":119,"rank":54,"notes":107,"score":120,"access":121,"effort":122,"family":123,"is_tier":108,"metrics":124,"cost_usd":107,"hardware":107,"model_id":43,"provider":107,"settings":125,"data_date":133,"num_cases":109,"precision":107,"score_std":107,"categories":134,"harness_id":107,"is_primary":108,"model_name":12,"model_slug":13,"output_tps":107,"short_name":12,"variant_id":144,"effort_rank":47,"input_price":107,"num_samples":72,"vendor_name":8,"vendor_slug":9,"architecture":107,"benchmark_id":47,"evaluated_at":116,"harness_name":107,"harness_slug":107,"input_tokens":145,"output_price":107,"overall_rank":72,"ranked_count":54,"release_date":107,"ttft_seconds":107,"variant_name":146,"variant_slug":147,"vendor_order":113,"weights_repo":107,"missing_count":113,"output_tokens":148,"overall_score":14,"parent_run_id":107,"context_window":107,"is_primary_run":108,"total_cost_usd":107,"total_params_b":107,"weights_source":107,"active_params_b":107,"cache_hit_price":107,"cost_incomplete":108,"harness_version":107,"is_open_weights":108,"avg_case_seconds":107,"inference_engine":107,"reasoning_tokens":107,"cost_per_case_usd":107,"trajectory_bundle_size_bytes":149},115,"0.2489","self_hosted","xhigh","Qwen3.8",{},{"top_p":126,"recipe":107,"harness":107,"judge_model":127,"num_samples":72,"temperature":128,"extra_settings":129,"scoring_method":107,"harness_version":107,"framework_version":131,"max_output_tokens":132},"0.950","DeepSeek-V4-Flash","0.60",{"top_k":130},-1,"OPT 9539216",106496,"2026-10-22",[135,138,140,142],{"name":41,"rank":136,"slug":42,"score":137,"expected":43,"available":43,"ranked_count":72},3,"48.453230452674897000000000",{"name":45,"rank":72,"slug":46,"score":139,"expected":47,"available":47,"ranked_count":72},"45.520000000000000000000000",{"name":49,"rank":43,"slug":50,"score":141,"expected":43,"available":43,"ranked_count":72},"49.480000000000000000000000",{"name":52,"rank":136,"slug":53,"score":143,"expected":54,"available":54,"ranked_count":72},"41.500000000000000000000000",19,88155012,"Qwen3.8-27B 自部署 xhigh","qwen3-8-27b-self-hosted-xhigh",100219144,191329325,[151],{"name":41,"slug":42},[153],{"name":90,"slug":91,"num_cases":109},{"extra_settings":155,"framework_version":157,"harness":158,"harness_version":159,"judge_model":160,"max_output_tokens":161,"num_samples":162,"recipe":163,"scoring_method":164,"temperature":165,"top_p":166},{"value":156,"varies":4,"label":107},{"top_k":130},{"value":131,"varies":4,"label":107},{"value":107,"varies":4,"label":107},{"value":107,"varies":4,"label":107},{"value":127,"varies":4,"label":107},{"value":132,"varies":4,"label":107},{"value":72,"varies":4,"label":107},{"value":107,"varies":4,"label":107},{"value":107,"varies":4,"label":107},{"value":128,"varies":4,"label":107},{"value":126,"varies":4,"label":107}]