Artificial Analysis升级编码代理指数至v2.1
Artificial Analysis表示,其编码代理指数v2.1以FrontierCode-AA 2.0取代FrontierCode v1.1,并将Terminal-Bench从4.0更新至4.1。新版指数中,Claude Fable 5.2与Claude Opus 5.5均以58分并列第一。Claude Fable 5.2在max配置下得56分,更多尝试触及任务时限。Artificial Analysis称,模型平均得分下降5分是FrontierCode-AA 2.0升级所致。在得分至少48的配置中,单任务成本从1.51美元到44.22美元不等。
Announcing Artificial Analysis Coding Agent Index v2.1, upgrading FrontierCode and Terminal-Bench. Claude Fable 5.2 and Claude Opus 5.5 share the lead at 58
The Coding Agent Index measures models in agent harnesses; v2.1 of the Index upgrades FrontierCode from 1.1 to 2.0 and Terminal-Bench from 4.0 to 4.1.
Upgraded:
➤ FrontierCode-AA 2.0 (replacing FrontierCode v1.1): our implementation of Cognition's FrontierCode 2.0, comprising 132 real-world coding tasks graded on whether a repository maintainer would merge the agent's change
➤ Terminal-Bench 4.1 (replacing Terminal-Bench 4.0): the same 66 tasks, scored with updated verifiers that improve reliability and resistance to reward hacking
Staying in the Index: SWE-Atlas-QnA-AA
#Artificial Analysis#FrontierCode-AA 2.0#Terminal-Bench 4.1#Claude Fable 5.2#Claude Opus 5.5#Claude Sonnet 5.5#Gemini 4 Carbon#GPT-6.1 Terra#GPT-6.1 Sol#Claude Code#Devin Fusion CLI#Codex