minirouter.sh
Open navigation
Get a key
Releases

Muse Glimmer 30B

Meta
MiniRouter and Muse Glimmer 30B by Meta, separated by a silver fold of light on black.

Meta's 30B open-weight dense multimodal model for local, long-running agentic workflows, distilled from the larger Muse Spark teacher and sized to run on a single consumer GPU.

Muse Glimmer 30B is available on MiniRouter as . The model page has current pricing, provider routes and API setup.

View Muse Glimmer 30B

About the model

  • Price: Not published. Meta positions it for local use with no per-token cost; hosted routes price independently
  • Weights: Released. Apache 2.0 on Hugging Face, in BF16, GGUF, and ExecuTorch builds
  • Parameters: 30B dense. ~29.6B per the model card, distilled from the Muse Spark teacher
  • Context: 128K. Model card lists 131,072+ tokens combined input and output

Publisher pricing

Pricing below comes from the linked publisher sources and does not represent a MiniRouter route or price.

Not published

The publisher announcement does not list an API price. MiniRouter does not estimate one.

See current MiniRouter rates and how billing works.

Publisher-reported benchmarks

Results from the publisher’s announcement, grouped by unit.

BenchmarkReported resultReported byAs ofSource
SWE-Bench Verified76.0MetaAug 10, 2026Announcement ↗
SWE-Bench Pro51.2MetaAug 10, 2026Announcement ↗
TerminalBench 2.151.7MetaAug 10, 2026Announcement ↗
OSWorld-Verified65.9MetaAug 10, 2026Announcement ↗
GPQA Diamond83.5MetaAug 10, 2026Announcement ↗

Meta's comparison baselines are Gemma4-31B and Qwen3.6-27B.

Meta notes ~4-bit quantization may reduce performance versus full precision.

Meta reports the benchmark figures; MiniRouter has not reproduced them.

Independent measurements

Artificial Analysis evaluations of the first-party API.

Muse Glimmer (high) ranks #136 of 462 model families on their Intelligence Index.

Intelligence Index
17.5
Composite of their evaluations
Coding Index
49.0
Coding evaluations only
Blended price
$0.64
USD per 1M tokens, 3:1 input to output, first-party list
Output speed
89
Median tokens per second
First answer token
23s
Median seconds, including reasoning

Muse Glimmer (high) on Artificial Analysis ↗

Source: Artificial Analysis ↗, read Sep 21, 2026. Measured on the model's first-party API, not a MiniRouter route.

Keep exploring

Compare published results in the model rankings and coding benchmarks, or read more model releases.

For MiniRouter product updates, visit the changelog.

Sources

  1. Meta AI Research announcement

    Published Aug 10, 2026

    Read at Meta ↗

  2. Artificial Analysis model page

    Published Sep 21, 2026

    Read at Artificial Analysis ↗

Editorial reference revised 2026-09-29. Catalog status comes from the generated routing snapshot.