Kilo now shows benchmark data right in the model picker, CLI and VS Code. Each model lists its Terminal Bench completion score and average cost per attempt, measured in Kilo's own harness. GPT-5.5 completes 74.1% of tasks, Kimi K2.6 hits 54.4%. Full table: kilo.ai/leaderboard

We ran GLM-5.2 and Kimi K2.7 Code through the same test: plan a feature flag service, then build it. GLM's plan scored 9.0 to Kimi's 8.1. But once both built from GLM's plan, the services were near identical. The planner matters more than the builder now. blog.kilo.ai/p/glm-52-vs-...

Inceptron is live in the Kilo Gateway. Sovereign, EU-hosted inference without giving up model choice. Kimi K2.6, GLM 5.1, MiniMax M2.5. GDPR and ISO 27001 compliant. From $0.15/1M input tokens on MiniMax M2.5. blog.kilo.ai/p/kilo-partn...

$15 on Opus. $1.70 on the same task with a cheaper model. Most of what an agent does all day doesn't need a frontier model, you're just paying like it does. Here's how to cut your AI bill: blog.kilo.ai/p/4-spend-le...

SpaceX is buying Cursor for $60 billion. SpaceX has the compute, Cursor has the distribution into half the Fortune 500, and the base models are converging. Once a tool gets acquired, its model choices serve the acquirer, not you. blog.kilo.ai/p/spacex-jus...

Fable 5 and Mythos 5 got pulled 3 days after launch over a US export directive. The frontier didn't go with them. GPT-5.5 tops KiloBench, Nemotron 3 Ultra is free and self-hostable, MiniMax M3 runs ~1/40th Fable's cost. The frontier is wider than one lab. blog.kilo.ai/p/you-dont-h...

Better than Fable 5, better than Le Chaton Fat, and better than whatever you're switching to tomorrow. Multiple models beats one single model. Every time.

The group chat will no longer be the one telling you your team lost. World Cup ClawByte, one-click install in KiloClaw, daily summaries + get yesterday's scores in your time zone, all piped to Telegram.

Dad what was it like when we had access to Fable?

ICYMI: Product Week shipped five things this week. All of them, in one place 🧵

Most AI review tools run the same generic rubric on every repo and miss what your team actually cares about. Code Reviews now adapt: drop a REVIEWS.md in your repo and the agent enforces your conventions, not someone's default. Live now.

Two new coding models and a sold-out token plan this week in Kilo. Kimi K2.7 Code from Moonshot, Claude Fable 5 topping our coding benchmarks, and the MiniMax token plans selling out fast enough to need a new batch already.

The Kilo CLI now has a face. Kilo Console is a local, browser-based UI for managing your projects, git worktrees, sessions, and settings. No more hand-editing JSON. Now in beta.

"Open doesn't just mean weights." Chris Alexiuk of NVIDIA on everything the Nemotron family opens up: the data, the recipes, the technical report. The whole point is an open science community around models.

A secret got committed to Git, then "erased" with a branch rewind. The repo looked clean. We ran Grok Build 0.1 on this Terminal-Bench task. It found the orphaned commit, saved the secret, and actually scrubbed the history. 27 steps, 41 seconds, $0.09. Writeup: blog.kilo.ai/p/we-asked-g...

MiniMax M3 benches near Claude Opus 4.8 at a tenth of the price. Coding Plans are live in Kilo. Buy them with the balance you already have, no separate subscription.

On June 1, GitHub Copilot switched to token-based billing. Developers are reporting bills 10x higher, with some burning through a month of credits in hours. Meanwhile, a quieter shift has been underway. There's something for switchers at the end of this thread. 🧵

The dentist appointment you forgot. The standup it knew you'd miss. The Telegram ping to fix it. This is KiloClaw now. 🦞

Hit by the June 1 Copilot billing change? Bring your ChatGPT Plus or Pro subscription into Kilo and run GPT-5.5 at a flat rate, in your IDE or CLI. kilo.ai/use-chatgpt-...

Same prompt, three models, three answers. You keep the one you like. Agent Manager is live in Kilo Code. Every agent runs in its own git worktree, so they work in parallel without touching each other's files.

3 points higher on the benchmark. 8x more expensive to actually run. Guess which one was the right model to ship? Benchmarks tell you who wins. They don't tell you what it costs to win. KiloBench does. blog.kilo.ai/p/kilobench-...

The lesson from the Copilot news last week goes past "find a cheaper vendor." The real exposure is being tied to one vendor's next price change at all. Cost control comes from transparency and model freedom: see what you'll pay, route each task to the right model. kilo.codes/v3IpnYc

We gave the same code audit to Claude Opus 4.8 and MiniMax M3. Same codebase, same prompt, 17 known bugs planted in advance. MiniMax M3 caught 13 for $0.07. The cheapest Claude run caught the same 13 for $1.30.Here's the breakdown. 🧵

Three new arrivals in Kilo this week, all pushing on context, speed, and cost. Kilo is already using more Nemotron 3 Ultra than Hermes and OpenClaw combined.

Five days. Five ships. 🚀 Next week is Product Week at Kilo. Agent Manager. KiloClaw. Console. Coding plans. Code reviews. One a day, Monday through Friday. Starts Monday. 👀

The GitHub Copilot bill came due June 1. But the scariest number this week isn't from Copilot. Uber burned its entire 2026 AI coding budget by April. Four months in. And it wasn't even Copilot. It was Claude Code and Cursor.

Two new open models. Free in Kilo. 👉 Nemotron 3 Ultra (NVIDIA): strongest US open-weights model. Agentic reasoning, 5x faster, 1M context. 👉 Step 3.7 Flash (StepFun): multimodal, 400 tok/sec, 256k context. Near-Sonnet coding at a fraction of the cost. $0. kilo.codes/BQHE4qE

Codebase indexing is back in Kilo Code. GA today. Semantic search for when you know what code does but not what this repo calls it. One query beats four guessed greps. Core implementation from community contributor shssoichiro. Opt-in & local-friendly. blog.kilo.ai/p/codebase-i...

Your AI planner takes a vague prompt, guesses what you meant, and hands you a confident plan built on assumptions you never agreed to. The new Architect agent refuses. It grills you first, then plans. Live now in the Kilo Marketplace: blog.kilo.ai/p/architect-...

Someone asked Claude Opus 4.8 "what model are you?" in Chinese. It said "I'm Qwen." The internet decided Anthropic distilled Qwen. Then it said DeepSeek. Then Claude. A real distillation fingerprint fails the same way every time. This one didn't. blog.kilo.ai/p/did-claude-opus-48-distill-alibabas

Roo Code shut down May 15th. 3M installs, 23K stars, gone. Same week Google pushed Antigravity 2.0 as a background update and people opened their IDE to find a chatbot staring back. The tool you built your workflow around isn't the tool you have anymore. kilo.codes/IEjacv3

"Middleman" is usually an insult. This is the good kind. Bring your own provider key! They still run the model and bill you. Kilo Gateway just adds the analytics and one-key-everywhere your provider never did. 0% markup, while OpenRouter and Requesty take 5%. kilo.codes/JUqGMfY

Claude Opus 4.8 is in Kilo Code, and we're running 20% off inference through our provider network. Frontier model, discounted price. BYOK or route through us, toggle the dropdown and let your agents loose. kilo.codes/Fp7TENu

Voting closes tonight on the May Product Hunt rankings! Here's one more kindness from this month. If v7 has been part of how you ship, your vote keeps it on top: www.producthunt.com/products/kil...

Best single-line review of v7 so far! Last day to vote for #1 Product of the Month: www.producthunt.com/products/kil...

Parallel agents was one of the bets in v7 that took the longest to get right. Then this showed up on Product Hunt. Two days left to vote: www.producthunt.com/products/kil...

Three new arrivals in Kilo this week, all sharpening the price-to-performance trade. Opus 4.8 at the same price as 4.7. Step 3.7 Flash from StepFun. And Xiaomi just cut MiMo pricing by up to 99%. kilo.codes/BQHE4qE

Kilo Code v7 has been #1 Product of the Month on Product Hunt all month. This is the kind of feedback that put us there. ❤️ Vote to keep us on top! www.producthunt.com/products/kil...

We gave Grok Build 0.1 one prompt: build a webhook delivery service in TypeScript, Bun, and SQLite.It planned it, built it, shipped a working demo.Total cost: $1.65. Zero tool-calling failures. Here's what happened, file by file. 🧵

Last week, Elon replied to a Grok Build 0.1 test. This time, he quote-tweeted one: "Good value for money." One prompt, a full webhook delivery service in TypeScript, Bun, and SQLite. Built and shipped for $1.65 total. Zero tool-calling failures. Follow along for more builds!

And it's already live in Kilo Code. Open the model dropdown, pick Opus 4.8, go.

Kilo Gateway now does Bring Your Own Key across 20 providers and 24 plans. DeepSeek, xAI, Anthropic, OpenAI, and more. We charge 0% on top. OpenRouter and Requesty both add 5%. One key works across every Kilo Code product area, plus a full usage dashboard. kilo.codes/o1oz0z1

Are parallel coding agents actually useful, or just online theater? We asked 7 of our senior engineers how they're really using them. Some run 20+ at once. Some cap it at 2. Here's what we found. 🧵 kilo.codes/0i9p5bv

If you're already paying for X Premium+ or SuperGrok, you shouldn't have to pay twice to use those models in your coding agent. Starting today, connect your Grok subscription directly to Kilo Code. No separate API key. Includes Grok Build 0.1 for agentic coding. kilo.codes/1DXA9XV

Claude Opus 4.7, Opus 4.6, and Sonnet 4.6 are 20% off through the Kilo Gateway right now. Ari ran the numbers: one power user saves ~$660/month. A team of 10 keeps $6,600. kilo.codes/bwGub4b

Brian built a 3D SpaceX Starship simulator in Kilo Code with Grok Build 0.1. Two prompts, 35 cents, under 5 minutes. Skill is open source: github.com/bturcotte520/holographic-renderings

Stepfun is still leading the Kilo Gateway with 30%+ of usage. But three new arrivals are shaking it up. • Grok Build 0.1 made some noise. • Gemini 3.5 Flash quietly raised the bar. • Qwen 3.7 Max showed up earlier than expected. Same workflow, same prompts, three different answers.

Round two on Grok Build 0.1. 7 new sites we built in Kilo Code: parrot translator, moon hotel booking, Pomodoro timer, CRT terminal, Shakespeare's GitHub, Tokyo weather, lemonade stand dashboard. Prompts, costs, and code analysis for each: kilo.codes/8rjlDPY

A professional-level landing page. 40 seconds. Less than ten cents. That's what a developer built with Grok Build 0.1, xAI's new coding model. Live in Kilo now. 256K context, no output limits, $1/$2 per million tokens. Already climbing the leaderboard. 📈 kilo.codes/0sWmPLC

Qwen3.7 Max is live in Kilo! Alibaba's new flagship for the agent era. 1M context, built for long-horizon coding tasks. Switch your model and see what it does.