DEV Community

Darko from Kilo
Darko from Kilo

Posted on

We hosted a Live Session on Gemini 3 Pro Preview. Here's how it went.

#gemini #llm #ai

We just wrapped an action-packed session walking through Google's new Gemini 3 model and how it performs inside Kilo Code. We talked benchmarks and pricing, and even shipped an entire video game with one prompt using Kilo Deploy.

Highlights

Headline: Gemini 3 is Google's most capable model so far.

Why should developers care?

Pricing vs Claude Sonnet 4.5

  • At typical prompt sizes (<200k tokens), Gemini 3 Pro is slightly cheaper per input/output token than Claude Sonnet 4.5.

  • In long-context (1M tokens), Gemini is more cost-efficient on both input and output.

  • With Kilo's open model selection, you could use Gemini for heavy reasoning/design, then switch to more inexpensive models for smaller implementation tasks, without changing anything about your workflow.

Live Demo: Building & Deploying Kilo Man in One Prompt

We showed an end-to-end flow using Gemini 3 Pro Preview in Kilo Code to build a playable (and honestly, pretty fun) video game in one prompt:

Prompting Kilo in Orchestrator Mode
Here's the prompt we gave to Kilo Code, set to Orchestrator mode:

"Create a modern, detailed 2D platform game called "Kilo Man" with a humanoid character that has to avoid obstacles and moving "monsters" to reach a final goal (like Super Mario).

It should be full screen, with a parallax, aesthetic background, and a black and yellow theme. The game camera should follow the player, and there should only be one, fairly long, scrolling level.

Include a slider that adjusts the player's jumping height."\
Here's what happened next:

Agent Orchestration in Kilo Code

-   Kilo's **[Orchestrator mode](https://kilo.ai/docs/basic-usage/orchestrator-mode)** created a subtask for **[Architect](https://kilo.ai/docs/basic-usage/using-modes#architect-mode)** mode.

-   **Architect** wrote an implementation and architecture plan to a markdown file in the project repository:

    -   Tech stack: **Next.js 15, React, Tailwind, HTML5 Canvas**

    -   Systems for camera, enemies, physics, parallax visuals, and a slider for jump height.

-   Orchestrator then handed off to **Code** mode, which implemented the full app according to the plan.        

    ![](https://substackcdn.com/image/fetch/$s_!25qJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff06df5cb-cce4-4906-9b3c-7c8cd49a428a_2802x1756.png)
Enter fullscreen mode Exit fullscreen mode

Result: a fully playable platformer

-   Fullscreen, parallax background, patrolling enemies, restart on death, follow-camera, jump-height slide. None of this was micromanaged in the prompt.

-   All generated in one shot, then ran locally:

    ![](https://substackcdn.com/image/fetch/$s_!EHSW!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9c893b77-513b-457b-a6b3-faa99873725f_3442x1764.png)
Enter fullscreen mode Exit fullscreen mode

Deploying with Kilo Deploy (early testing feature)

-   We connected GitHub to Kilo via the **[Integrations](https://app.kilocode.ai/integrations)**[ tab ](https://app.kilocode.ai/integrations)in the Kilo Dashboard.

-   Went to **Deploy** → "New Deployment" → selected the Kilo Man repo & branch → clicked **Create deployment**.

-   Kilo built the project and gave a **shareable production URL** for the game:

    ![](https://substackcdn.com/image/fetch/$s_!BwF0!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F33ff3284-6e8f-461b-9711-9065a2484585_2456x1726.png)
Enter fullscreen mode Exit fullscreen mode

You can play Kilo Man here:

And, you can be an early tester of Kilo Deploy today! Visit our Discord to learn more:

Questions We Covered (Q&A Highlights)

  • Q: What's Kilo Code's long-term plan for autonomous debugging?

    • Today: Use our native Debug mode, which you can pair with any model (including Gemini 3). Also, you can use the CLI in autonomous mode with the --auto flag.
    • Coming soon: Sign up to be an early tester of code reviews and cloud agents!

  • Q: How does Kilo Code ship so fast?

    • Internally, we call it "Kilo Speed". We have a tight team, fast shipping cadence, and quick feedback loop with our open source community.

  • Q: Is the Kilo CLI actively developed and core to the product?

  • Q: Is it better to access Gemini 3 via Kilo Gateway or CLI? How does tool calling work?

    • It's the same model either way, so you can choose the CLI or Extension based on your personal workflow preference.

  • Q: Do parallel agents use Git worktrees under the hood?

    • Yes, they do! That's how they don't overwrite/conflict with each other.

  • Q: Is Kilo based on Cline/Roo and is it up to date?

    • Kilo originally forked Roo, then pulled in all the best from Cline.
    • Kilo is a superset of the best open-source features: take the best from what already exists, and keep evolving from there.

  • Q: Are you accepting community contributions? Any best practices?

    • Absolutely!
    • Contributions, bug reports, and feature requests go through GitHub Discussions, so feel free to post there with details and screenshots, if possible.

  • Q: Any more events coming?

    • Yes! more online events are being planned.
    • To stay tuned, join our Discord and Reddit channels.

What This All Means

  • Gemini 3 brings frontier-level reasoning + huge context + good taste in code.

  • Kilo Code's agentic orchestration (Architect / Code / Ask / Debug / Orchestrator) lets you actually use that intelligence in real projects.

  • With Kilo Deploy, going from "idea in a prompt" → "live app on a URL" is now realistically a one-prompt + few-clicks experience.

And, watch the session replay here:

Replay the Gemini 3 Webinar

Stay tuned for more events!

Top comments (0)