This measures the amount of verifiable public evidence we have, not how capable the model is. A missing field means it has not been verified yet, not that its value is zero.
We’ve written an FAQ to answer some of the questions we've received about watermarking. In summary: • We’re implementing watermarking to comply with the EU AI Act. Other major model developers have si
We’ve written an FAQ to answer some of the questions we've received about watermarking. In summary: • We’re implementing watermarking to comply with the EU AI Act. Other major model developers have signed the same Code of Practice and will also be implementing watermarking;
Claude Security scans now run on Claude Mythos 5, available today in public beta for all Claude Enterprise customers. Put our most capable security model to work on your codebase, no separate model ac
Claude Security scans now run on Claude Mythos 5, available today in public beta for all Claude Enterprise customers. Put our most capable security model to work on your codebase, no separate model access needed. https://t.co/zJSUgGjtZF
LongWoF-Bench: Evaluating EvoMap Genes for Verifiable Long-Workflow Tasks
Large language models are increasingly expected to execute complex workflows whose success depends on maintaining interdependent constraints and producing artifacts that satisfy strict end-to-end verification. Yet successful execution experience is typically lost after a single run, forcing subsequent models to rediscover strategies and failure modes from scratch. We study whether such experience can instead be externalized and reused through EvoMap, where verifier-confirmed execution trajectories are consolidated into structured Gene. To evaluate this setting, we introduce the Long-Workflow Benchmark (LongWoF-Bench), comprising 778 machine-verifiable tasks across code generation, agent-environment synthesis, mathematical reasoning, and rule following. On the 252 tasks with verifier-confirmed Opus trajectories, evolved EvoMap Gene outperform Skill across all seven evaluated models by 8.7-15.5 percentage points, with the gains extending to consumer models from different model families. In contrast, reference-distilled Gene do not exhibit the same advantage, indicating that compact representation alone is insufficient and that Gene utility is closely associated with verified experience provenance. For Claude Opus, Gene reuse also completes 39 more tasks than Skill while reducing solve-time token consumption by 9.9%. Together, these results show that verified execution experience can be retained and shared as a reusable external resource, enabling models to improve long-workflow completion without repeatedly paying the full cost of experience discovery.
SWE-bench Science: Can Coding Agents Resolve Engineering Tasks in Science?
Software increasingly functions as part of the scientific instrument itself, making failures in scientific code capable of compromising not only program behavior but also the evidence underlying scientific conclusions. Yet existing evaluations of coding agents largely emphasize aggregate task success, providing limited insight into why agents fail when repairing scientific software. We introduce SWE-bench Science, a repository-level benchmark for scientific software engineering comprising 119 tasks from 98 GitHub repositories across 20 scientific domains. Each task is organized into one of three paradigms: Issue-driven, Expert-exploratory, and Engineering-integration. Even the best-performing agent, Claude Code with Opus-5 (max), achieves a pass@1 below 50\%, highlighting the substantial challenges posed by scientific software engineering. We identify four recurring failure mechanisms: deficits in scientific knowledge or abstraction, misguided exploration or surface-level repair, incomplete repair coverage or system integration, and failures to generalize scientific knowledge beyond observed cases in our analysis. We further conduct a paired ablation that removes explicit scientific guidance while preserving the repository and executable engineering context. The results show that scientific knowledge is not uniformly beneficial: well-grounded information can constrain repair and improve average performance and token efficiency, whereas poorly aligned guidance can induce anchoring and does not necessarily improve exact repair success. Together, SWE-bench Science provides a broad testbed for studying both the capabilities and failure mechanisms of coding agents in scientific software engineering.
Claude Security scans now run on Claude Mythos 5, available today in public beta for all Claude Enterprise customers. Put our most capable security model to work on your codebase, no separate model ac
Claude Security scans now run on Claude Mythos 5, available today in public beta for all Claude Enterprise customers. Put our most capable security model to work on your codebase, no separate model access needed. https://t.co/zJSUgGjtZF
X/Twitter@claudeaiAnthropicannouncementgeneral4d ago
Claude Academy is now live. Whether you're figuring out what AI is or already using Claude every day, there's a path that meets you where you are. The courses and tutorials are free and open to anyone
Claude Academy is now live. Whether you're figuring out what AI is or already using Claude every day, there's a path that meets you where you are. The courses and tutorials are free and open to anyone at https://t.co/WRiSRAvK6l https://t.co/llx2W0VIIy
X/Twitter@claudeaiAnthropicannouncementgeneral6d ago
Claude can now send emails in Gmail and manage files in Google Drive. Ask Claude to reply to a thread, and it drafts and sends the response. You control when it needs your approval. Connect Gmail or G
Claude can now send emails in Gmail and manage files in Google Drive. Ask Claude to reply to a thread, and it drafts and sends the response. You control when it needs your approval. Connect Gmail or Google Drive from the connectors menu to try. Available on all paid plans. https://t.co/cFZEjh3MgB
We’ve written an FAQ to answer some of the questions we've received about watermarking. In summary: • We’re implementing watermarking to comply with the EU AI Act. Other major model developers have si
We’ve written an FAQ to answer some of the questions we've received about watermarking. In summary: • We’re implementing watermarking to comply with the EU AI Act. Other major model developers have signed the same Code of Practice and will also be implementing watermarking;
LongWoF-Bench: Evaluating EvoMap Genes for Verifiable Long-Workflow Tasks
Large language models are increasingly expected to execute complex workflows whose success depends on maintaining interdependent constraints and producing artifacts that satisfy strict end-to-end verification. Yet successful execution experience is typically lost after a single run, forcing subsequent models to rediscover strategies and failure modes from scratch. We study whether such experience can instead be externalized and reused through EvoMap, where verifier-confirmed execution trajectories are consolidated into structured Gene. To evaluate this setting, we introduce the Long-Workflow Benchmark (LongWoF-Bench), comprising 778 machine-verifiable tasks across code generation, agent-environment synthesis, mathematical reasoning, and rule following. On the 252 tasks with verifier-confirmed Opus trajectories, evolved EvoMap Gene outperform Skill across all seven evaluated models by 8.7-15.5 percentage points, with the gains extending to consumer models from different model families. In contrast, reference-distilled Gene do not exhibit the same advantage, indicating that compact representation alone is insufficient and that Gene utility is closely associated with verified experience provenance. For Claude Opus, Gene reuse also completes 39 more tasks than Skill while reducing solve-time token consumption by 9.9%. Together, these results show that verified execution experience can be retained and shared as a reusable external resource, enabling models to improve long-workflow completion without repeatedly paying the full cost of experience discovery.
SWE-bench Science: Can Coding Agents Resolve Engineering Tasks in Science?
Software increasingly functions as part of the scientific instrument itself, making failures in scientific code capable of compromising not only program behavior but also the evidence underlying scientific conclusions. Yet existing evaluations of coding agents largely emphasize aggregate task success, providing limited insight into why agents fail when repairing scientific software. We introduce SWE-bench Science, a repository-level benchmark for scientific software engineering comprising 119 tasks from 98 GitHub repositories across 20 scientific domains. Each task is organized into one of three paradigms: Issue-driven, Expert-exploratory, and Engineering-integration. Even the best-performing agent, Claude Code with Opus-5 (max), achieves a pass@1 below 50\%, highlighting the substantial challenges posed by scientific software engineering. We identify four recurring failure mechanisms: deficits in scientific knowledge or abstraction, misguided exploration or surface-level repair, incomplete repair coverage or system integration, and failures to generalize scientific knowledge beyond observed cases in our analysis. We further conduct a paired ablation that removes explicit scientific guidance while preserving the repository and executable engineering context. The results show that scientific knowledge is not uniformly beneficial: well-grounded information can constrain repair and improve average performance and token efficiency, whereas poorly aligned guidance can induce anchoring and does not necessarily improve exact repair success. Together, SWE-bench Science provides a broad testbed for studying both the capabilities and failure mechanisms of coding agents in scientific software engineering.
ClawGym II: Exploring Black-Box RL on Agent Harness
Agent harnesses have substantially improved performance on long-horizon tasks by coordinating agent interactions with the environment. However, reinforcement learning through complex harnesses remains largely unexplored, as scaling such training to long-horizon agent tasks introduces fundamental challenges. In this work, we present a unified black-box RL framework for stable and scalable optimization of general agents through complex harnesses. Concretely, we first build a sandbox-based execution infrastructure that isolates task environments and harnesses within temporary sandboxes for large-scale concurrent rollouts. We then decouple policy optimization from opaque harness execution and place a serving proxy at the model boundary to capture model calls. To reconstruct multi-turn trajectories and improve training efficiency, we organize the captured calls into prefix trees and further adapt both critic-based PPO and critic-free GRPO to optimize over the recovered tree structure. Meanwhile, we maintain training-inference consistency throughout the optimization process. Finally, we introduce mix-harness training, allowing a single model to be jointly optimized by heterogeneous harnesses. With Qwen3-30A3B, black-box RL improves Pass@1 on ClawGym-Bench by 9.98 and 14.81 points through OpenClaw and Claude Code, respectively, while remaining stable over 200-400 optimization steps. Moreover, the framework yields consistent gains on more challenging tasks such as JobBench and OfficeQA. Overall, our framework enables effective, stable, and scalable optimization of general agents through black-box harnesses, supporting unified training across heterogeneous execution systems.
