Alibaba’s open-source terminal coding agent Qwen Code shipped v0.24.6 on 2026-09-26 (npm @qwen-code/qwen-code 0.24.6; Desktop desktop-v0.24.6 and SDK TypeScript sdk-typescript-v0.1.16 same day): CLI declares Managed Runtime v2 execute/status/cancel and mounts tool ops on the worker; adds a native Advisor consultation tool with usage limits; `/batch-api` for agent-prepared batch workflows; serve gains guarded Hosted Runtime, Spring control plane + dual-path WebShell, and model/reasoning selection at session creation; core adds Durable Managed Session journal and failover. Release notes list no known breaking changes.
Key Takeaways
- ✓Shipped: v0.24.6 / npm 0.24.6; Desktop desktop-v0.24.6 same window
- ✓Runtime: Managed Runtime v2 execute/status/cancel contract; tool ops mounted on the worker
- ✓Advisor: native advisor tool with consultation behavior and usage limits
- ✓Batch: /batch-api agent-prepared Batch API workflow
- ✓Hosted: guarded Hosted Runtime + Durable Managed Session journal/failover; model/reasoning at session create; no known breaking changes
Project Links & Resources
Direct AccessIn-Depth Technical Analysis
Core Background & Industry Pain Points
Terminal coding agents need standardized execution contracts, tool lifecycles, and durable sessions to move beyond local chat into hosted, batch, and failover-ready deployments. Qwen Code’s Managed/Hosted paths still split execute/status/cancel, tool receipts, and long-session recovery; teams lacked a first-class Advisor surface and Batch API entry, and Java/SDK vs CLI worker bindings could drift.
Architecture Highlights & Internals
v0.24.6 (2026-09-26) declares Managed Runtime v2 execute/status/cancel, mounts v2 tool ops on the worker, and extends sdk-java with Hosted Harness client, W0a Managed Workspace binding, and Runtime-evidence reconciliation for UNKNOWN tool executions. Adds a native advisor tool with usage limits and [/batch-api](https://github.com/QwenLM/qwen-code/pull/12492). Serve gains guarded Hosted Runtime, Spring control plane + dual-path WebShell; core adds Durable Managed Session journal/failover and model/reasoning at session create. Companion Desktop and sdk-typescript-v0.1.16.
Authoritative Benchmarks & Measured Scores
No public SWE-Bench/latency scores in the notes; a startup benchmark harness (#12674) lands for local regressions. Versus v0.24.5 the delta is Runtime v2, Advisor, /batch-api, and Durable Session—not a model-score jump.
Developer Hands-on Guide
npm i -g @qwen-code/[email protected] or follow the user overview; verify qwen sessions ps, try /batch-api and Advisor, and configure Hosted Runtime via Alibaba Cloud docs. Release: GitHub.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.