Are AI coding agents actually getting better? Six months of diagnostics data say yes, with caveats.
Kiro’s agents use LSP-based diagnostics to catch type errors, unresolved imports, and undefined symbols. Newer models aren’t just making fewer errors, they’re checking broader
System prompt behavior is hard to predict.
An instruction that looks fine in isolation can break in cases nobody tested for.
We built an LLM-judge pipeline to catch it before it ships.
27 candidates screened. Kiro CLI behavioral quality issues down 32%.
Full breakdown 👉
GPT‑5.6 🤝 Kiro
GPT‑5.6, now available in Kiro, advances price-performance for developers.
Result: 82% cost reduction per successful task, in under 4 minutes, at 74% pass rate on Terminal-Bench 2.1. Engineering rigor, quality, and intelligence.
Details here:
Kiro office hours just moved to Mondays.
10:30pm PT, live on Discord. Get your questions answered in real time.
First one’s today 👉 http://spr.ly/6013B1s1k5
Kiro changed how Kavya, a Solutions Architect at AWS, thinks about her own time.
Hours turned into minutes, without trading away accuracy. Her latest build: a React-based LLM summarizer, shipped in 30 minutes.
More room to experiment, less time on grunt work.
One session, any surface.
Cloud sessions are live in Kiro CLI and IDE.
Start a task from the CLI, close the laptop, come back to it done. Or fire off a longer-running task from the browser, pick up in the terminal or IDE when you're ready to dig in.
Which Kiro app should you pick? 👀
We’re building a family of front doors onto one increasingly capable engine, each shaped for a different person and a different moment, none of them trying to eat the others.
Pick the door that fits the task:
•IDE → code-first work
•CLI →
Before Kiro, Howard Fan spent most of his time on delivery.
Now the Lead AI Engineer at T. Rowe Price spends on product thinking, how to help the business, not just ship projects.
Try it today 👉 http://spr.ly/6010B1MmTg
Kiro powers now support Agent Plugins. ⚡
No more packaging the same skill three different ways for three different clients. Publish once to the open specification (built by AWS, Cursor, Microsoft, OpenAI, Vercel), install anywhere the standard is supported. Existing powers
Get to know Kiro Crew from the engineers who built it 🔥
Kiro Crew is an open source, persistent development workspace that remembers your context, lessons, and skills across sessions, so you come back to progress instead of a cold start.
Join us to hear the origin story
6/ Secure by design: 7-layer defense-in-depth, from OS-level sandboxing to signed audit logs. Because it’s open source, you can verify every layer yourself.
GitHub → http://github.com/kirodotdev/kirocrew
5/ Apps: purpose-built interfaces for work that doesn’t belong in chat. Issue Triage, Research Assistant, Agent Manager, and more, ship at launch, or build your own.
4/ Multi-agent orchestration: run several conversations in parallel, or delegate work to subagents that report back. Nothing waits in line.
Docs → http://kiro.dev/docs/crew
3/ It works when you don’t. Multi-step tasks execute unattended, recurring jobs run on your schedule, and heartbeats monitor your systems and deployments until something needs you.
2/ Work from anywhere: start on the desktop app, web dashboard, or CLI. Continue in Slack, Discord, Telegram, Teams, Webex, WeCom, or WeChat. Context carries across all of them.
Aggregated from the public X timeline; copyright belongs to the original author. This site is not affiliated with this account.
💬 Need a hand?
Purchase / payment / account help — chat with us →
Support online
Support online 9:00–21:00 (UTC+8), we reply promptly
Leave your email so we can follow up and pick up where we left off next time.
Announcements
If you have a credit card, you can register an account on this site, use your credit card to top up your balance in the personal center, and then use the balance to pay for purchasing our products or services.
