Encoder V1
Meta🇺🇸Frontier
Embeddingopen
MetaEncoder v1 produces embeddings for images, video, documents and text that suit search, clustering and comparison. Its decide function also answers closed-set questions about a state with probabilities.
This week2 models
Muse Ember
Meta🇺🇸Frontier
TextReasoningVisionAPI
Muse Ember is the smallest and most efficient model from Meta that’s designed for complex reasoning execution and the work agents repeat all day: tool calls, coding agent subtasks, routing,...
Week of October 12, 20262 models
Muse Spark Open
Meta🇺🇸Frontier
TextReasoningVisionopenAPI
Meta coding model for agentic workflows with 816B total and 42B active MoE parameters, a 1M token context window, and text, image, video and audio input. NVFP4 checkpoints are offered for download on Hugging Face.
Muse Wonder
Meta🇺🇸Frontier
TextReasoningVisionAPI
Meta announces Muse Wonder and Muse Wonder Full, which it calls its most capable models for personal and professional agentic tasks such as insurance claims, leases, and building apps. Muse Wonder is available via the Meta Model API, Muse Wonder Full only to Muse Code subscribers.
Week of September 28, 20261 model
Encoder
Meta🇺🇸Frontier
Embeddingopen
No description yet — the model card or announcement hasn't been summarised.
Week of August 31, 20261 model
Muse Spark 1.3
Meta🇺🇸Frontier
TextReasoningVisionAPI
Muse Spark 1.3 Contributor is the cost-efficient contributor tier of Meta’s multimodal reasoning model for experimentation, learning, and early-stage agentic, multi-agent, and coding workflows. It is designed to track information...
1 variantContributor
Week of August 3, 20262 models
Muse Glimmer
Meta🇺🇸Frontier
TextReasoningVisionopenAPI
A dense 30B open-weight model from Meta for local agentic work, covering multi-step reasoning, tool use and image understanding. According to the sources it runs fully on a single GPU without cloud access.
5 variants30B30B Executorch Pte30B Assistant30B GGUF30B
Muse Spark 1.2
Meta🇺🇸Frontier
TextReasoningVisionAPI
Muse Spark 1.2 contributor tier is a reasoning model from Meta designed for developers who want to start building at an even lower cost. It’s meaningfully cheaper than Muse Spark...
1 variantContributor
Week of July 13, 20261 model
Muse Spark 1.1
Meta🇺🇸Frontier
TextReasoningVisionAPI
Muse Spark 1.1 is a multimodal reasoning model from Meta, built for agentic tasks. It accepts text, images, video, audio, and PDF documents and returns text, with a 1M-token context...
Week of June 8, 20261 model
MobileMoE-S
Meta🇺🇸Frontier
Textopen
No description yet — the model card or announcement hasn't been summarised.
2 variantsBaseSFT
Week of March 31, 20251 model
Llama 4 Maverick
Meta🇺🇸Frontier
Visionopen
Meta's multimodal MoE flagship: 17B active, 128 experts, 402B total, 1M-token context. Scout variant (16E, 109B) for lean deployments. Llama 5 is expected in H2 2026.