DataLearner logo
MU

Muse Glimmer-30B

Reasoning modelTool use

Muse Glimmer-30B by Meta Superintelligence Labs

Release date: 2026-08-10Knowledge cutoff: 2026-01-045
Live demoGitHubHugging FaceCompare
Parameters
29.6B
Context length
128K+
Chinese support
Not supported
Reasoning ability

Muse Glimmer-30B by Meta Superintelligence Labs is an AI model published by Facebook AI Research Lab, released on 2026-08-10, for Reasoning model, with 29.6B parameters, and 128K+ context length, with a 953.00 score on GDPval-AA v2.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Muse Glimmer-30B

Model basics

Reasoning traces
Supported
Thinking modes
Thinking Level · LowThinking Level · MediumThinking Level · HighThinking Level · Extra-High
Context length
128K+ tokens
Max output length
No data
Model type
Reasoning model
Modality (in / out)
Text, Image → Text
Release date
2026-08-10
Model file size
No data
MoE architecture
No
Total params / Active params
29.6B / N/A
Knowledge cutoff
2026-01-04
Muse Glimmer-30B

Open source & experience

Code license
Weights license
Apache 2.0- Commercial use permitted
GitHub repo
GitHub link unavailable
Live demo
No live demo
Muse Glimmer-30B

Official resources

Muse Glimmer-30B

API details

API speed
3/5
No public API pricing yet.
Muse Glimmer-30B

Benchmark Results

Muse Glimmer-30B currently shows benchmark results led by AA-LCR (1 / 16, score 80), IF Bench (4 / 31, score 77), AIME 2026 (6 / 19, score 94.70). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage
Internet

General Knowledge

2 evaluations
Benchmark / mode
Score
Rank/total
83.50
63 / 188
HLE
High
22
109 / 173

Coding and Software Engineer

3 evaluations
Benchmark / mode
Score
Rank/total
76
37 / 113
51.20
39 / 55
43.60
2 / 2

Instruction Following

1 evaluations
Benchmark / mode
Score
Rank/total
77
4 / 31

Long Context

1 evaluations
Benchmark / mode
Score
Rank/total
AA-LCR
High
80
1 / 16

Math and Reasoning

1 evaluations
Benchmark / mode
Score
Rank/total
94.70
6 / 19

AI Agent - Tool Usage

3 evaluations
Benchmark / mode
Score
Rank/total
MCP-Atlas
HighTools
75.50
12 / 28
65.90
19 / 25
51.70
29 / 29

Productivity Knowledge

1 evaluations
Benchmark / mode
Score
Rank/total
GDPval-AA v2
HighTools
953
6 / 6

AI Agent - Information Search

1 evaluations
Benchmark / mode
Score
Rank/total
DeepSearchQA
HighToolsInternet
74.60
2 / 2

Multimodal Understanding

3 evaluations
Benchmark / mode
Score
Rank/total
78.80
9 / 9
75.80
2 / 2
74
5 / 7

Agent Level Benchmark

1 evaluations
Benchmark / mode
Score
Rank/total
τ³-Banking
HighTools
23.50
2 / 2

Compare with other models

No curated comparisons for this model yet.

Want a custom combination? Open the compare tool

Muse Glimmer-30B

Publisher

Facebook AI Research Lab
Facebook AI Research Lab
View publisher details
Muse Glimmer-30B by Meta Superintelligence Labs

Model Overview

Muse Glimmer-30B by Meta Superintelligence Labs is an AI model published by Facebook AI Research Lab, released on 2026-08-10, for Reasoning model, with 29.6B parameters, and 128K+ context length, with a 953.00 score on GDPval-AA v2.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code