Muse Glimmer 30B is available on EmpirioLabs. Meta's open 30B agentic model runs on EmpirioLabs native inference with a 131,072-token context, text and image input, function calling, strict structured output, and adjustable reasoning strength. It is built for agent scaffolds, coding, and multi-step tool work.
What Muse Glimmer 30B supports
Muse Glimmer 30B accepts text and images and writes text. The perception encoder reads screenshots, charts, and documents alongside conversation, which makes it a strong fit for agents that look at real interfaces. Function tools can be called in parallel, and strict JSON Schema structured output returns exactly the shape you request.
Reasoning strength is adjustable. Use reasoning_effort from none through max to trade speed against depth; high is the default and max maps to the model's strongest setting. Reasoning text streams on the reasoning_content channel and reasoning tokens are reported in usage.
How to call Muse Glimmer 30B
Use the OpenAI-compatible chat completions endpoint and set model to muse-glimmer-30b:
curl https://api.empiriolabs.ai/v1/chat/completions \
-H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \
-H 'Content-Type: application/json' \
-d '{
"model": "muse-glimmer-30b",
"reasoning_effort": "high",
"messages": [
{
"role": "user",
"content": "Plan a staged rollout for a breaking API change."
}
]
}'The same model is available through the Responses, Messages, legacy completions, and Google-compatible content-generation endpoints. Streaming responses include final usage data, and cached input is reported so repeated prompts settle at the discounted cache rate.
Vision, tools, and structured output
Attach images as standard image_url content parts, including several images in one message. For machine-readable results, use JSON mode or provide a strict JSON Schema. Optional web search powered by Linkup can retrieve recent sources for your latest message and adds a small per-call charge when invoked.
Pricing
Muse Glimmer 30B uses pay-as-you-go token pricing with a steeply discounted rate for cached input and no per-seat fees. Current rates are on the model page and pricing.
Try it
Open the playground, read the API docs, or browse all models and pricing.
