Argus 这个智能体运行时，SWE-Bench Pro 拿到 78%，比 Direct Copilot 高 19 个点，还会自我进化省 token，挺新鲜。
Argus 是一个自进化的智能体运行时，由 Manager、Planner、Engineer、Reviewer 四个角色在持久项目状态上执行任务。在 SWE-Bench Pro 上达到约 78% 的通过率，高于 Direct Copilot 的 59%，但 token 消耗为后者的 1.41 倍。经过验证门控自进化后，成熟阶段的求解 token 比启动阶段少 21%，活动时间少 15%，并记录了 34 次验证器恢复和 22 次审查循环救援。在 AARRI-Bench 上取得 76.8%，数学数据合成差距为 28 分；优化后的 RWKV6 内核已合并到上游。
Argus: A General-Purpose Agentic Runtime for Long-Horizon Reasoning
Long-horizon reasoning requires an agentic runtime that can persist when evidence supports its current approach and pivot when measurements reveal failure, hidden constraints, or a misspecified objective. We present Argus, a persistent, self-evolving runtime in which Manager, Planner, Engineer, and Reviewer execute bounded missions over durable project state. Argus separates stable user intent from operational objectives, constraints, and verification criteria, and admits memories, skills, procedures, verifiers, routing decisions, and rejected routes only after role-owned review and, when available, task-native verification. Model weights remain fixed; self-evolution occurs through persistent runtime state and control policy, with autonomous execution between operator-owned escalation points. Across seven GPT-5.5 benchmark arenas, Argus achieves about 78% on SWE-Bench Pro versus 59% for Direct Copilot while using 1.41 times the aggregate tokens. After verification-gated self-evolution, mature SWE-Bench waves use 21% fewer solve-input tokens and 15% less active workflow time per task than startup waves, while recording 34 verifier recoveries and 22 strict review-loop rescues. Argus also reaches 76.8% on AARRI-Bench and a 28.0-point gap on mathematical data synthesis, with competitive GPU-kernel and language-model-training results. Beyond benchmarks, an optimized RWKV6 kernel was merged upstream; a multi-day mathematics campaign retained falsified routes and proof-backed frontier updates; and six paper pipelines completed 254 missions with 16 stage rollbacks. These results show that a fixed-weight, self-evolving harness can revise, recover, and accumulate verified approaches while producing structured trajectories for future supervised and reinforcement learning.