Editing Openai/691a14bb-ac48-8008-85d0-2e4fb3f780d1 (section)

===== What it is: =====
* A newer coding-focused model variant with a very large context window (~1 million tokens) according to public write-ups. Cline<ref>{{cite web|title=Cline|url=https://cline.ghost.io/code-supernova-1-million/|publisher=cline.ghost.io|access-date=2025-11-17}}</ref>
* Built for agentic coding (multimodal support like diagrams/screenshots, coding workflows) via platforms such as Kilo Code, etc. Cline<ref>{{cite web|title=Cline|url=https://cline.bot/blog/code-supernova-stealth-model|publisher=cline.bot|access-date=2025-11-17}}</ref>
* There are notes this model may no longer be widely available (“stealth” release, shut-down in some platforms). blog.kilocode.ai<ref>{{cite web|title=blog.kilocode.ai|url=https://blog.kilocode.ai/p/code-supernova-is-shutting-down-heres|publisher=blog.kilocode.ai|access-date=2025-11-17}}</ref>

Strengths:
* Massive context window: great when you need to feed in large amounts of code, history, docs, PRs and have the model reason across them.
* Tailored for coding workflows: good for many files, complex tasks, possibly image/diagram input.
* Could be a “go-to” when context size bottleneck is the issue.

Weaknesses / trade-offs:
* Availability may be limited (might not be fully supported in your environment).
* Possibly less mature or less tested in production (given the “stealth” status).
* Might still be slower or more expensive than more standard models.
* If your task doesn’t require massive context, the large window is wasted overhead.

When to use in Windsurf Cascade:
* Use when you want to feed in a very large set of code/documentation: e.g., full codebase + history + design docs + tests + many modules.
* Use when you want to integrate diagrams/screenshots + code + natural-language reasoning (if supported).
* Avoid if your task is small or you don’t need the huge context window — use a lighter/cheaper model instead.