Code Supernova Explodes: Now with a 1 Million Token Context Window!
Our current stealth frontier model just leveled up to compete with GPT4.1 and Claude Sonnet 4, while still being completely free in Kilo Code!
TL;DR: Code Supernova upgraded today (Sept 26) to a massive 1 million token context window putting it in the same league as the other top tier models out there like OpenAI’s GPT4.1 and Anthropic’s Claude Sonnet 4. Most important of all, it’s still completely free through Kilo Code; our 100% free, open source AI agent extension for VS Code and JetBrains with over 420,000 installs.
How It Started
When we first launched Code Supernova in stealth last week (Sept 19) it packed impressive multimodal capabilities into just a 200K context window. Today, they’re taking it to the next level.
In your settings, `code-supernova` will be listed as `code-supernova-1-million` and you’ll be able to test out the difference immediately!
Why It Matters
For folks who don’t know what context window means in this context (😏); you can think of it as your model’s “attention span”. The phrase “working memory” is also used quite often to describe this but it’s essentially just how much information your model can keep in mind while helping you code. A larger context window means you can share more files, longer conversations, and bigger projects without the AI forgetting what you talked about earlier.
It’s almost like having RAM in the cloud, which makes my pranks in college where I convinced my friends that they could download RAM online a little ironic!
So like I mentioned above, a 1 million token context window puts Code Supernova in elite company. For comparison:
GPT4.1: 1M tokens ($15/million tokens)
Claude Sonnet 4: 1M tokens ($15/million tokens)
Gemini 2.5 Pro: 1M tokens ($7/million tokens)
Code Supernova: 1M tokens ($0 through Kilo Code)
With Code Supernova, you now can:
Upload entire codebases (8+ copies of the full React repo!)
Add screenshots to those codebases for additional context
Ship a full product starting with just a design mockup or architecture diagram!
What You Should Do
Starting today:
Install Kilo Code
Download our VS Code extension or JetBrains plugin and join 420,000+ developers already using Kilo.
Select Code Supernova 1 million
Click on the model name under the prompt box and select Code Supernova 1 million from the list:
If you don’t see it pop up in the list, make sure you’re using Kilo Code as your API provider in the Settings!
Start Building!
Get this new frontier AI model at zero cost (for a limited time). Start building: Upload your files, screenshots, and documentation; all at once!
No rate limits. No token counting. No usage caps.
Ready to experience 1 million token context? Download Kilo Code and join 420,000+ developers already building with the latest in frontier AI!
 | A guest post by
|
I would like to express my sincere thanks to the developers of Kilo Code. Without their support, I would not have been able to move forward with my project VALEO-NeuroERP-2.0 nearly as fast as I did.
VALEO-NeuroERP is an open-source, next-generation ERP system built on a modular service-oriented architecture (MSOA) with deep AI integration. It is designed to be:
• Self-optimizing – continuous AI-assisted process optimization
• Voice-enabled – natural language interaction and command execution
• Predictive – advanced offer and order management driven by forecasting models
• Modular & scalable – CRM, Finance, Inventory, Logistics, Compliance, and more
• Cloud-native – Kubernetes-based, with CI/CD and full observability
• Agentic by design – LangGraph/MCP agents for automated coding, testing, and operations
The goal: a self-learning ERP platform that empowers businesses to run smarter, faster, and more efficiently — tailored for agricultural trade, yet adaptable to many industries.
Kilo Code and Supernova are on a roll! Amazing to see.