Documentation Navigation

Models & pricing / Models / Muse-Glimmer-30B

Muse-Glimmer-30B Overview

This model's data has not been fully verified and benchmarks may be missing.

Muse-Glimmer-30B is a multimodal-general AI model created by meta-models, featuring 29.8B parameters and a context window of unknown.

Muse Glimmer is a 30-billion-parameter causal language model with a dedicated perception 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 single model that runs locally without requiring cloud infrastructure or network access.

Model Lineage & Specification

muse-glimmer-30b is meta-models's primary release in the current family.

Developermeta-models
API Identifiermuse-glimmer-30b
Parameters29.8B
Context Windowunknown
LicenseUnknown

Comparable models

Muse-Glimmer-30B

Muse Glimmer is a 30-billion-parameter causal language model with a dedicated pe...

API IDmuse-glimmer-30b
Typeopen weights
Parameters29.8B
Contextunknown

Muse-Glimmer-30B-GGUF

Muse-Glimmer-30B-GGUF is a undisclosed-parameter image text to text model develo...

API IDmuse-glimmer-30b-gguf
Typeopen weights
Parameters29.6B
Context131,072+

Minimax-H3-nvfp4-INT4-INT8-Convrot

Minimax-H3-nvfp4-INT4-INT8-Convrot is a undisclosed-parameter image text to vide...

API IDminimax-h3-nvfp4-int4-int8-convrot
Typeopen weights
Parameters
Contextunknown

BigBang-v1

SGLang is a high-performance serving framework for large language models and mul...

API IDbigbang-v1
Typeopen weights
Parameters36.0B
Contextunknown

Muse Glimmer is a 30-billion-parameter causal language model with a dedicated perception 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 single model that runs locally without requiring cloud infrastructure or network access.