meta-models/Muse-Glimmer-30B

• Type: Generation
• Capabilities: vision, reasoning
• Cache read: $0.01 per 1M input tokens (0.1× standard input price). See prompt caching.

Overview

Muse Glimmer is a 30-billion-parameter language model with a dedicated image encoder, distilled from Muse Spark and purpose-built for autonomous agentic tasks on consumer hardware. The model integrates multi-step reasoning, reliable tool use, multimodal understanding, and failure recovery into a small model.

---

Reasoning: This model controls reasoning via prompting, not request parameters. Reasoning strength can be defined as part of the system prompt as Reasoning strength: . Muse Glimmer supports the following levels: low / medium / high / xhigh. Use high or xhigh for complex problem solving, coding, and agentic tasks.

Reasoning efforts

• Supported: none, minimal, low, medium, high, xhigh, max

See the reasoning effort guide for request examples.

Pricing

Priority	Input Tokens (per 1M)	Cache Read Tokens (per 1M)	Output Tokens (per 1M)
Realtime1	$0.11	$0.01	$0.32
Async	$0.07	$0.01	$0.24
Batch (24h)	$0.05	$0.01	$0.15

Playground

Open this model in the Playground.

Footnotes

1. Realtime availability is limited. Doubleword is primarily a batch API. ↩