Saltar al contenido principal

Muse Glimmer-30B

MetaOpen WeightApache 2.0 · Uso Comercial

Descripción

Muse Glimmer-30B is Meta Superintelligence Labs' open-weight, dense multimodal model for autonomous agentic work on consumer hardware. Its 29.6-billion-parameter causal transformer includes a roughly 1.8B-parameter ViT-G/14 perception encoder, accepts interleaved text and images, produces text, and supports a 131,072-token context window. It combines controllable multi-step reasoning, schema-based tool use, coding, failure recovery, and compatibility with agent scaffolds such as OpenClaw and Hermes Agent. Meta also releases 4-bit variants sized for 24 GB and 32 GB systems and a DFlash speculative-decoding drafter. The model is trained on more than 100 languages and released under Apache 2.0.

Fecha de lanzamiento
2026-08-10
Parámetros
29.6B
Longitud del contexto
131K
Modalidades
image, text

Radar de capacidades

60
general
60
coding
50
reasoning
51
scienceest.
50
agents
70
multimodal

Science se estima a partir de las puntuaciones científicas de LLM Stats o del razonamiento cuando no hay benchmarks científicos dedicados.

Rankings

Dominio#PosiciónPuntuaciónFuente
Capacidad agéntica69
44.0
LS
Ranking multimodal99
47.0
LS

Puntuaciones de benchmarks (LLM Stats)

(LLM Stats (zeroeval))

Agents

Siren AgentDojo Utility94.2%Aut.
CI Memories Coverage64.8%Aut.
WildClawBench47.6%Aut.
Siren AgentDojo Attack Success Rate28.4%Aut.
CI Memories Violation Rate26.4%Aut.
Tau3 Banking23.5%Aut.

Code

SkillsBench44.3%Aut.

General

GAIA243.3%Aut.

Instruction Following

IFBench77.0%Aut.

Math

AIME 202694.7%Aut.

Multimodal

OSWorld-Verified65.9%Aut.

Reasoning

GPQANYU + Cohere + Anthropic (2023)83.5%Aut.
AA-LCR80.0%Aut.
CharXiv-R78.8%Aut.
SWE-Bench Verified76.0%Aut.
MCP Atlas75.5%Aut.
Beam 128K65.1%Aut.
Terminal-Bench 2.151.7%Aut.
SWE-Bench ProPrinceton NLP (2024)51.2%Aut.
SciCode43.6%Aut.
Humanity's Last Exam22.0%Aut.

Search

DeepSearchQA74.6%Aut.

Vision

OmniDocBench 1.575.8%Aut.
ScreenSpot Pro75.4%Aut.
MMMU-Pro74.0%Aut.

Índices de evaluación AA

(Artificial Analysis)

No hay datos de evaluación AA disponibles

Puntuaciones por categoría LLM Stats

(LLM Stats (zeroeval))
Instruction Following
80
Long Context
80
Spatial Reasoning
80
Structured Output
80
Frontend Development
80
Grounding
80
Multimodal
70
Search
70
Vision
70
Physics
60
Reasoning
60
General
60
Biology
60
Chemistry
60
Code
60
Math
50
Agents
50
Coding
50
Tool Calling
50

Precios

Precio de entrada$0 / 1M tokens
Precio de salida$0 / 1M tokens
Precio mixto (3:1)$0 / 1M tokens
Precio de lectura caché$0.04 / 1M tokens

Velocidad

No hay datos de velocidad disponibles

Ranking de Precios por Proveedor

Ranking de Precios por Proveedor

8 proveedores

Más barato: MetaMás caro: Vercel AI Gateway
ProveedorEntradaSalida
1MetaPRINCIPAL
$0
$0
2DeepInfra
$0
$0
3EmpirioLabs AI
$0.2
$0.8
4OpenRouter
$0.3
$1.2
5Kilo Gateway
$0.3
$1.1
6DevPass (LLM Gateway)
$0.3
$1.2
7NanoGPT
$0.35
$1.5
8Vercel AI Gateway
$0.35
$1.5

Comparar precios entre diferentes proveedores de API para este modelo.

Fuentes externas