模型
场景榜单里出现过的模型。分数来自基准结果，不在这里单独维护。
模型排名价格 $/M上下文
Claude Opus 5.5WebDev #1$4 / $201M闭源GPT 5.6 LunaSlop Index #1——闭源GPT-6 AstraDeepSWE #1$10 / $501.1M闭源Claude Sonnet 5.5Image-to-WebDev #2$2 / $101M闭源Gemini 3.8 FlashDeepSWE #2——闭源Claude Fable 5.1Android Bench #3$10 / $501M闭源Claude Opus 5DeepSWE #3$5 / $251M闭源GPT 5.6 SolSlop Index #3——闭源GPT-6.1 SolTerminal-Bench #3$2 / $101.1M闭源DeepSeek V4.1 FlashSlop Index #4——MITGPT 6 SolAndroid Bench #4$2 / $101.1M闭源Claude Fable 5DeepSWE #5——闭源Qwen3.8 MaxAndroid Bench #5$2 / $61M闭源Gemini 3.1 Pro PreviewSlop Index #6——闭源GPT 5.6 TerraDeepSWE #6——闭源Kimi K3Android Bench #6$3 / $151MKimi K3 licenseGLM-5.3sudo L7 #7$1.40 / $4.40—MITGPT 6 LunaAndroid Bench #7——闭源Muse Spark 1.3Image-to-WebDev #8$1.25 / $4.251M闭源Gemini 4 ArgonWebDev #9$2 / $101M闭源Grok 4.6DeepSWE #9——闭源Claude Sonnet 5Android Bench #10——闭源HY4 Previewsudo L7 #10$0.83 / $2.501MApache 2.0GPT 5.5DeepSWE #11——闭源Grok 4.7sudo L7 #11$2 / $6500K闭源Gemini 3.7 FlashDeepSWE #12——闭源GLM 5.3 FlashDeepSWE #13——闭源DeepSeek V4 ProDeepSWE #14——闭源Claude Opus 4.8DeepSWE #15——闭源MiMo V2.6 FlashWebDev #16$0.14 / $0.281.1MMITMuse Spark 1.2DeepSWE #17——闭源Qwen3.8 Flash NextWebDev #17$0.16 / $0.471Mqwen-communityMiMo V2.6 ProImage-to-WebDev #18$0.43 / $0.871.1MMITGrok 4.5DeepSWE #19——闭源DeepSeek V4 FlashDeepSWE #20——闭源Muse Spark 1.1DeepSWE #21——闭源GPT 5.4DeepSWE #22——闭源Gemini 3.6 FlashDeepSWE #23——闭源GLM 5.2DeepSWE #24——闭源Gemini 3.5 FlashDeepSWE #25——闭源MiniMax M3Slop Index #25——闭源Kimi K2.7 CodeDeepSWE #26——闭源Claude Sonnet 4.6DeepSWE #27——闭源Claude Haiku 5.5WebDev #30——闭源Cursor Composer 2.5———闭源