Can't Beat Free: Gemma 4 31B on AMD Token Lab, with a Local Chat Page and an ADK Agent
AMD Token Lab is a free OpenAI-compatible API on AMD GPUs that comes with the AMD AI Developer Program. How to sign up, a local chat page that keeps the key off the browser, a terminal client, and a small ADK agent that calls Gemma 4 31B through it.
First seen by Oriel on .
AI summary
This step-by-step guide demonstrates building a local chat page, terminal client, and Google ADK agent using Gemma 4 31B through AMD Token Lab’s free API.
- AMD AI Developer Program members receive 10 points, 30 days of access, and 20 requests per minute without a credit request or payment.
- The local standard-library Python chat page proxies requests, keeps the API key off the browser, supports images and thinking, and displays timing and usage.
- A Google ADK agent connects through one LiteLlm object and uses tools for exact counting, range calculations, and model-service listings.
- Across 23 requests, the examples consumed 0.000335 points, while Gemma 4 31B supported text, images, tools, JSON output, and a 262,144-token context.
- Shared capacity caused variable performance: first tokens arrived after 1.2–5.0 seconds, decoding reached 2.0–9.5 tokens per second, and overload produced HTTP 429 responses.
#token lab #gemma #sglang #google adk
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Reactions
Total public reactions reported by the source; distinct from points and likes.
- DEV Community: 40, latest observation ; +8 over observed period since
Comments
Comment count reported by the source.
- DEV Community: 13, latest observation ; 0 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| DEV Community | 5031066 | https://dev.to/gde/cant-beat-free-gemma-4-31b-on-amd-token-lab-with-a-local-chat-page-and-an-adk-agent-c5b Read discussion |