Termsslug: subagent

Subagent

子智能体

Independent AI agent modules executing specific sub-tasks under main agent coordination

Curated Intelligence

8 items · Page 1/1
GREG ISENBERG (@gregisenberg)T2·Models
78 pts
Anthropic Ships Claude Fable and Mythos 5.2 Model Pair

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2, claiming gains in coding, reasoning and vision. Early observations report 10-12 hour autonomous work sessions, a jump from 21% to 41% on real freelance tasks in a single release, and orchestration of 1,000+ subagents. If accurate, this marks a step change in long-horizon agent autonomy and large-scale multi-agent coordination.

Hacker News AIT2·Agents
78 pts
Octomind Launches: Open-Source Rust AI Agent Runtime with MCP Support

Octomind is released as a Rust-based, CLI-first AI coding agent runtime that grants LLMs real access to repositories via MCP tools for file operations and shell execution. It supports multiple modes including interactive, piped stdin, background daemons, and WebSocket, featuring built-in hooks and validators to replace manual approvals in autonomous workflows.

Hacker News AIT2·Agents
78 pts
Post-mortem: Claude Multi-Agent System Ports F-Zero X to 3DS

A developer released a post-mortem on porting G-Diffuser (an F-Zero X decompilation) to the Nintendo 3DS using AI agents. Orchestrated by a single long-lived Claude session directing 147 subagents, the project achieved 60fps stereoscopic gameplay on New 3DS in 24 days, later expanding to Old 3DS support and native rendering. The effort consumed ~3.17 billion tokens with an API cost of nearly $5,000, demonstrating the feasibility and economics of large-scale autonomous coding agents for complex legacy system migrations.

InfoQ 中文T2·Agents
78 pts
Google Cloud Launches Gemini Agent: Introduces 'Coworker' Identity and Multi-Model Orchestration

Google Cloud unveiled Gemini Agent at the Gemini at Work 2026 event, introducing the concept of 'Coworker Agents' that possess independent enterprise identities including email, calendar, and persistent storage. The agent executes cross-application tasks via a single prompt, supports sub-agent collaboration, and breaks model lock-in by currently integrating Anthropic Claude with plans for broader multi-model support. It is currently in private preview for enterprise knowledge work scenarios.

AWS Machine Learning BlogT1·Models
45 pts
Claude Haiku 5.5 Launches on AWS: 75% Cheaper for High-Volume Tasks

Anthropic announced the availability of Claude Haiku 5.5 on Amazon Bedrock and Claude Platform on AWS. Positioned as the fastest and most efficient model in the Claude 5.5 family, it targets subagents and high-volume workloads with costs approximately 75% lower than Haiku 4.5. The integration offers regional data residency and unified AWS billing while maintaining native platform capabilities.

宝玉 (@dotey)T1·Models
86 pts
Anthropic Launches Claude Haiku 5.5: Input Price Drops 90%, OSWorld Score Jumps to 72.4%

Anthropic released Claude Haiku 5.5, reducing short-context input prices by 90% and average running costs by ~75%. The model's score on the OSWorld 2.1 benchmark surged from 15.7% to 72.4%, and it introduces a five-level 'effort' control for cost-performance balancing. Additionally, Sonnet 5.5 cache read prices were halved, and Max/Team subscribers now receive monthly API credits.

LangChain BlogT1·Agents
68 pts
LangChain Open-Sources Deep Life Sci: Multi-Agent Framework for Life Sciences with Literature Retrieval and Secure Data Analysis

LangChain has released Deep Life Sci, an open-source multi-agent assistant designed for clinical and lab scientists. By integrating tens of millions of scientific articles, performing data analysis in a secure sandbox, and delegating tasks to hundreds of sub-agents, it addresses the lack of domain-specific knowledge in general AI tools. The framework also allows R&D teams to inject private data and utilize LangSmith for tracing and evaluation, ensuring transparency and customization in research workflows.