Kilo Code

07/28/2026

Six major model releases in the last month alone. Keeping track of which one is best (and cheapest) for each task is now a job on its own.

We keep arriving at the same conclusion in these experiments: you don't need the most expensive model for every step of your workflow. So we tested whether Kilo Code's Auto Model router could act on that better than picking models by hand.

All three built the same working service, but Kilo's Auto Model built it for less than half the cost:

- Auto Model (frontier plan → efficient build): $1.26
- GPT-5.6 Sol everywhere: $2.90
- Claude Sonnet 5 everywhere: $2.47

Once a good plan exists, the model that builds it moves the cost far more than the result.

The pace of new model releases has made “which model should I use for this?” a real problem.

07/23/2026

More ways to control your AI coding spend are live in Kilo

We just shipped part 2 of our cost optimization features:

📊 Cost Insights: see exactly what's driving your bill, broken down by product and member

🔔 Spend Alerts: get notified before an unusual hour turns into a surprise invoice

💡 Cost Suggestions: we now actively point you to whichever inference path (coding plan, Kilo Pass, and more) fits your usage best

Read the full post: blog.kilo.ai/more-ways-to-control-ai-coding-spend

07/23/2026

Everyone's waiting for a second "DeepSeek moment." We don't think it's coming.

Open models like Kimi K3 top the intelligence charts, but serving them made throughput fall off a cliff.

The scarce thing was never the model, but the compute to run it.

Read more: https://blog.kilo.ai/p/no-second-deekseek-moment

07/10/2026

Kilo Code is now available for JetBrains IDEs as a native Kotlin/Swing plugin built on the IntelliJ Platform. It includes chat, slash commands, file mentions, MCP servers, and model selection. Available now on the JetBrains Marketplace.

A ground-up rebuild in Kotlin brings first-class AI coding assistance to IntelliJ, WebStorm, PyCharm, and every JetBrains IDE

06/30/2026

Auto Efficient is a next-generation model router inside Kilo. It's session-aware and informed by real benchmark data.

It reads your prompts and requests, consults KiloBench data to find the cheapest model proven to handle that type of task, then automatically routes your request there.

As your coding session progresses, Auto Efficient learns and adjusts its routing strategy to optimize for cost and performance.

On average, it's 77% cheaper than Claude Opus 4.8 while retaining nearly 70% performance parity.

Check out the numbers here: https://kilo.ai/auto-efficient-vs-opus

06/24/2026

Let’s take a minute to acknowledge the “wait, have we talked already?” problem. 😅

When you’re managing a lot of relationships, it’s easy to lose track of who got followed up with, when you last talked, and what it was about. 🤯

KiloClaw builds a personal CRM automatically. Every day it reads your emails and calendar activity and updates a memory file for each contact. So before you reach out, just ask “what do I know about this person?” and get back the last conversation, topics discussed, last touch point, and what to follow up on.

No more guessing. No more inbox digging. 🙌

Try KiloClaw for free at kilo.ai (link in bio).

06/23/2026

You’re paying frontier prices to rename a variable. 💸

Auto Efficient routes each request to the cheapest model that can handle it, picked on a public benchmark you can check. Easy tasks run lean, hard ones stay reliable.

It reads the session as it happens, works out how hard the task is, and routes accordingly. You don’t switch anything. If a routing call isn’t clear-cut, it falls back to a capable paid model, so quality has a floor.

Live now in the model picker. ⚡

06/19/2026

“Sometimes we just need a hammer. We don’t need a multi-tool.” 🔨

Chris Alexiuk of on a shift a lot of people are sleeping on: if a model only ever needs to do one thing, it doesn’t need to be good at ten others. Make it much better at the one thing that matters, or much smaller, which usually means cheaper and faster. ⚡

Generalists still matter. But like people, models specialize too.

06/17/2026

Most people skip meeting prep because they don’t have time. 📅

KiloClaw watches the calendar. When it spots a first-time external meeting in the next 30 minutes, it researches who you’re meeting with, pulls their background, their company, recent news, and talking points, and drops a briefing right into Slack. ⚡

Show up informed. No lifting a finger. 🙌

Try KiloClaw for free! Link in bio.