{"title":"Publisher-reported coding model benchmarks","revisedAt":"2026-08-14","methodology":{"source":"Publisher announcements","independent":false,"comparableHarnesses":false,"overallRank":null},"models":[{"slug":"gemini-3-7-flash","name":"Gemini 3.7 Flash","vendor":"Google","announcedAt":"2026-08-13","inMiniRouterCatalog":false,"pricePhases":[{"label":"Introductory","inputPerMillionUsd":0.75,"outputPerMillionUsd":3.75,"blendedPerMillionUsd":1.5,"effectiveFrom":"2026-08-13","effectiveTo":"2026-12-31","sourceUrl":"https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/"},{"label":"Standard","inputPerMillionUsd":1.5,"outputPerMillionUsd":7.5,"blendedPerMillionUsd":3,"effectiveFrom":"2027-01-01","effectiveTo":null,"sourceUrl":"https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/"}],"benchmarks":[{"id":"frontiercode-1-1","name":"FrontierCode 1.1","value":43.6,"displayValue":"43.6%","unit":"percent","reportedBy":"Google","sourceDate":"2026-08-13","sourceUrl":"https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/","harnessVerifiedByMiniRouter":false},{"id":"deepswe-1-1","name":"DeepSWE v1.1","value":65.3,"displayValue":"65.3%","unit":"percent","reportedBy":"Google","sourceDate":"2026-08-13","sourceUrl":"https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/","harnessVerifiedByMiniRouter":false},{"id":"webdev-arena","name":"WebDev Arena","value":1588,"displayValue":"1588 Elo","unit":"elo","reportedBy":"Google","sourceDate":"2026-08-13","sourceUrl":"https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/","harnessVerifiedByMiniRouter":false}]},{"slug":"glm-5-3","name":"GLM-5.3","vendor":"Z.ai","announcedAt":"2026-08-14","inMiniRouterCatalog":false,"pricePhases":null,"benchmarks":[{"id":"terminal-bench-3-0","name":"Terminal Bench 3.0","value":28.3,"displayValue":"28.3%","unit":"percent","reportedBy":"Z.ai","sourceDate":"2026-08-14","sourceUrl":"https://z.ai/blog/glm-5.3","harnessVerifiedByMiniRouter":false},{"id":"deepswe-1-1","name":"DeepSWE v1.1","value":66.9,"displayValue":"66.9%","unit":"percent","reportedBy":"Z.ai","sourceDate":"2026-08-14","sourceUrl":"https://z.ai/blog/glm-5.3","harnessVerifiedByMiniRouter":false},{"id":"cybergym","name":"CyberGym","value":84.5,"displayValue":"84.5%","unit":"percent","reportedBy":"Z.ai","sourceDate":"2026-08-14","sourceUrl":"https://z.ai/blog/glm-5.3","harnessVerifiedByMiniRouter":false},{"id":"exploitbench","name":"ExploitBench","value":54.4,"displayValue":"54.4%","unit":"percent","reportedBy":"Z.ai","sourceDate":"2026-08-14","sourceUrl":"https://z.ai/blog/glm-5.3","harnessVerifiedByMiniRouter":false},{"id":"exploitgym-2h","name":"ExploitGym · 2 hours","value":105,"displayValue":"105 tasks","unit":"count","reportedBy":"Z.ai","sourceDate":"2026-08-14","sourceUrl":"https://z.ai/blog/glm-5.3","harnessVerifiedByMiniRouter":false},{"id":"exploitgym-6h","name":"ExploitGym · 6 hours","value":130,"displayValue":"130 tasks","unit":"count","reportedBy":"Z.ai","sourceDate":"2026-08-14","sourceUrl":"https://z.ai/blog/glm-5.3","harnessVerifiedByMiniRouter":false}]},{"slug":"deepseek-v4-pro-0813","name":"DeepSeek-V4-Pro","vendor":"DeepSeek","announcedAt":"2026-08-13","inMiniRouterCatalog":false,"pricePhases":[{"label":"Introductory","inputPerMillionUsd":0.435,"outputPerMillionUsd":0.87,"blendedPerMillionUsd":0.54375,"effectiveFrom":"2026-08-13","effectiveTo":"2026-08-16","sourceUrl":"https://api-docs.deepseek.com/news/news260813/"},{"label":"Standard · peak","inputPerMillionUsd":1.32,"outputPerMillionUsd":3.96,"blendedPerMillionUsd":1.98,"effectiveFrom":"2026-08-16","effectiveTo":null,"sourceUrl":"https://api-docs.deepseek.com/news/news260813/"}],"benchmarks":[{"id":"terminal-bench-2-1","name":"Terminal Bench 2.1","value":87.9,"displayValue":"87.9","unit":"percent","reportedBy":"DeepSeek","sourceDate":"2026-08-13","sourceUrl":"https://api-docs.deepseek.com/news/news260813/","harnessVerifiedByMiniRouter":false},{"id":"hle-with-tools","name":"HLE (with tools)","value":60,"displayValue":"60.0","unit":"percent","reportedBy":"DeepSeek","sourceDate":"2026-08-13","sourceUrl":"https://api-docs.deepseek.com/news/news260813/","harnessVerifiedByMiniRouter":false},{"id":"deepswe","name":"DeepSWE","value":62.7,"displayValue":"62.7","unit":"percent","reportedBy":"DeepSeek","sourceDate":"2026-08-13","sourceUrl":"https://api-docs.deepseek.com/news/news260813/","harnessVerifiedByMiniRouter":false},{"id":"cybergym","name":"Cybergym","value":83.3,"displayValue":"83.3","unit":"percent","reportedBy":"DeepSeek","sourceDate":"2026-08-13","sourceUrl":"https://api-docs.deepseek.com/news/news260813/","harnessVerifiedByMiniRouter":false},{"id":"toolathlon-verified","name":"Toolathlon-Verified","value":74.1,"displayValue":"74.1","unit":"percent","reportedBy":"DeepSeek","sourceDate":"2026-08-13","sourceUrl":"https://api-docs.deepseek.com/news/news260813/","harnessVerifiedByMiniRouter":false}]},{"slug":"grok-4-6","name":"Grok 4.6","vendor":"xAI","announcedAt":"2026-08-12","inMiniRouterCatalog":false,"pricePhases":[{"label":"Standard","inputPerMillionUsd":2,"outputPerMillionUsd":6,"blendedPerMillionUsd":3,"effectiveFrom":"2026-08-12","effectiveTo":null,"sourceUrl":"https://x.ai/news/grok-4-6"}],"benchmarks":[{"id":"deepswe-1-1","name":"DeepSWE v1.1","value":65.9,"displayValue":"65.9%","unit":"percent","reportedBy":"xAI","sourceDate":"2026-08-12","sourceUrl":"https://x.ai/news/grok-4-6","harnessVerifiedByMiniRouter":false},{"id":"frontiercode-1-1-extended","name":"FrontierCode v1.1 (Extended)","value":61.3,"displayValue":"61.3%","unit":"percent","reportedBy":"xAI","sourceDate":"2026-08-12","sourceUrl":"https://x.ai/news/grok-4-6","harnessVerifiedByMiniRouter":false},{"id":"cursorbench-3-2","name":"CursorBench v3.2","value":69.9,"displayValue":"69.9%","unit":"percent","reportedBy":"xAI","sourceDate":"2026-08-12","sourceUrl":"https://x.ai/news/grok-4-6","harnessVerifiedByMiniRouter":false},{"id":"apex-swe","name":"APEX-SWE","value":56.4,"displayValue":"56.4%","unit":"percent","reportedBy":"xAI","sourceDate":"2026-08-12","sourceUrl":"https://x.ai/news/grok-4-6","harnessVerifiedByMiniRouter":false},{"id":"terminal-bench-3-0","name":"Terminal-Bench v3.0","value":26,"displayValue":"26%","unit":"percent","reportedBy":"xAI","sourceDate":"2026-08-12","sourceUrl":"https://x.ai/news/grok-4-6","harnessVerifiedByMiniRouter":false},{"id":"gdpval-aa-2","name":"GDPVal-AA v2","value":1753,"displayValue":"1753","unit":"elo","reportedBy":"xAI","sourceDate":"2026-08-12","sourceUrl":"https://x.ai/news/grok-4-6","harnessVerifiedByMiniRouter":false}]},{"slug":"muse-glimmer-30b","name":"Muse Glimmer 30B","vendor":"Meta","announcedAt":"2026-08-10","inMiniRouterCatalog":false,"pricePhases":null,"benchmarks":[{"id":"swe-bench-verified","name":"SWE-Bench Verified","value":76,"displayValue":"76.0","unit":"percent","reportedBy":"Meta","sourceDate":"2026-08-10","sourceUrl":"https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model","harnessVerifiedByMiniRouter":false},{"id":"swe-bench-pro","name":"SWE-Bench Pro","value":51.2,"displayValue":"51.2","unit":"percent","reportedBy":"Meta","sourceDate":"2026-08-10","sourceUrl":"https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model","harnessVerifiedByMiniRouter":false},{"id":"terminal-bench-2-1","name":"TerminalBench 2.1","value":51.7,"displayValue":"51.7","unit":"percent","reportedBy":"Meta","sourceDate":"2026-08-10","sourceUrl":"https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model","harnessVerifiedByMiniRouter":false},{"id":"osworld-verified","name":"OSWorld-Verified","value":65.9,"displayValue":"65.9","unit":"percent","reportedBy":"Meta","sourceDate":"2026-08-10","sourceUrl":"https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model","harnessVerifiedByMiniRouter":false},{"id":"gpqa-diamond","name":"GPQA Diamond","value":83.5,"displayValue":"83.5","unit":"percent","reportedBy":"Meta","sourceDate":"2026-08-10","sourceUrl":"https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model","harnessVerifiedByMiniRouter":false}]},{"slug":"qwen3-8-2-4t-a95b","name":"Qwen3.8-2.4T-A95B","vendor":"Alibaba","announcedAt":"2026-08-03","inMiniRouterCatalog":false,"pricePhases":[{"label":"Model Studio API","inputPerMillionUsd":1.65,"outputPerMillionUsd":4.951,"blendedPerMillionUsd":2.47525,"effectiveFrom":"2026-08-03","effectiveTo":null,"sourceUrl":"https://www.alibabacloud.com/blog/603421"}],"benchmarks":[{"id":"terminal-bench-2-1","name":"Terminal Bench 2.1","value":86.6,"displayValue":"86.6","unit":"percent","reportedBy":"Alibaba Cloud","sourceDate":"2026-08-03","sourceUrl":"https://www.alibabacloud.com/blog/603421","harnessVerifiedByMiniRouter":false},{"id":"swe-bench-pro","name":"SWE-bench Pro","value":67.7,"displayValue":"67.7","unit":"percent","reportedBy":"Alibaba Cloud","sourceDate":"2026-08-03","sourceUrl":"https://www.alibabacloud.com/blog/603421","harnessVerifiedByMiniRouter":false},{"id":"deepswe-1-1","name":"DeepSWE 1.1","value":56.6,"displayValue":"56.6","unit":"percent","reportedBy":"Alibaba Cloud","sourceDate":"2026-08-03","sourceUrl":"https://www.alibabacloud.com/blog/603421","harnessVerifiedByMiniRouter":false},{"id":"gpqa-diamond","name":"GPQA Diamond","value":92.6,"displayValue":"92.6","unit":"percent","reportedBy":"Alibaba Cloud","sourceDate":"2026-08-03","sourceUrl":"https://www.alibabacloud.com/blog/603421","harnessVerifiedByMiniRouter":false},{"id":"hle","name":"HLE","value":43.6,"displayValue":"43.6","unit":"percent","reportedBy":"Alibaba Cloud","sourceDate":"2026-08-03","sourceUrl":"https://www.alibabacloud.com/blog/603421","harnessVerifiedByMiniRouter":false}]}]}