Meta

Muse Glimmer 30B

by Metatext+image->text4 endpoints
Neura Intelligence Index
55.9/ 100
Rank
#160
Confidence
high
Benchmarks
9(23% coverage)

Overview

Muse Glimmer 30B is a dense, open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for autonomous agents on consumer hardware. It is suited for long-horizon agentic and coding workflows, with multi-step reasoning, reliable tool use, failure recovery, image understanding, and multilingual support across more than 100 languages.

Capabilities

Text generation
Image understanding
Tool use / Function calling
Structured output
Extended reasoning
Prompt caching
Long context
Top-P sampling
Stop sequences
Deterministic seed
Log probabilities

Modalities

Input
TextImage
Output
Text

Technical Specifications

Context Window
131.1K tokens
Max Output
118.0K tokens
Tokenizer
Other
Uptime (24h)
10000.0%
Parameters
29600000000B
License
apache_2_0Open Weights
Country
🇺🇸US
Release Date
August 10, 20261mo

Supported Parameters

frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p

Benchmark Performance

coding

BenchmarkScoreSource
SWE-bench Verified
76
verified
SWE-Bench Pro
51.2
verified
SciCode
43.6
verified

reasoning

BenchmarkScoreSource
GPQA Diamond
83.5
verified
Humanity's Last Exam
22
verified

multimodal

BenchmarkScoreSource
CharXiv Reasoning
78.8
verified
ScreenSpot Pro
75.4
verified
MMMU-Pro
74
verified

agent

BenchmarkScoreSource
MCP Atlas
75.5
verified

Performance Over Time

Pricing

Per 1M tokens
Input$0.30
Output$1.10
Cache read$0.04
Added
August 9, 2026
Last synced 9/18/2026