Tag
An article comparing two agentic software engineering benchmarks — Frontier Code v1.1 and Terminal Bench v4 — and questioning why Kimi K3 scores well on one while performing poorly on the other.
Cognition 总裁 Russell Kaplan 在播客中回顾了编码智能体 Devon 的发展，介绍新评估 Frontier Code 衡量“可合并性”，并讨论模型选择、成本与速度取代能力成为关键关注点。
In the interview, Cognition President Russell Kaplan reviews the development of coding agent Devon, discusses model intelligence saturation and cost optimization, and introduces how the new evaluation Frontier Code measures code mergeability.
Claude Opus 5 is now available in Devin, offering near-Fable-level performance on FrontierCode 1.1 at half the cost.
Cognition releases FrontierCode 1.1, an updated benchmark for evaluating code quality with refined guidelines for fair internet use and grading criteria, along with new model scores for Sonnet 5 and Fable 5.
Claude Sonnet 5, a new AI model with frontier-level coding performance at a more affordable price, is now available in Devin Desktop and Devin CLI, outperforming Opus 4.8 on FrontierCode Extended.