Meta Superintelligence Labs
PulseAugur coverage of Meta Superintelligence Labs — every cluster mentioning Meta Superintelligence Labs across labs, papers, and developer communities, ranked by signal.
- developed by Muse Image 95%
- developed Muse Wonder 95%
- developed Muse Image 95%
- instance of Muse Spark 90%
- developed by Muse Spark 90%
- developed Meta Model API 90%
- used by Meta Model API 90%
- developed Muse Glimmer 90%
- affiliated with WhatsApp 90%
- employed by Alexandr Wang 90%
- developed Muse Code 90%
- developed Muse Spark 90%
- 2026-10-13 product_launch Meta Superintelligence Labs launched its second-generation model family, Muse Wonder. source
7 day(s) with sentiment data
Meta's strategic shift to private APIs for frontier models like Muse Spark contrasts with past open-weight leadership
Recent evidence indicates Meta has moved away from releasing open-weight models, with its latest frontier model, Muse Spark, only available as a private API. This marks a significant departure from its previous role in the open-weight LLM community, suggesting a strategic pivot towards proprietary AI development and monetization.
Meta to release Muse Video API within 90 days, leveraging Muse Image advancements
Meta has announced Muse Video as a preview model, sharing a pretraining base with Muse Image. Given the successful rollout of Muse Image and its integration across Meta platforms, it's plausible Meta will prioritize a public API for Muse Video to capitalize on its generative video capabilities, likely within the next three months.
Muse Image's integration with social context may lead to new forms of AI-driven social media content creation and moderation challenges
Muse Image's ability to incorporate 'social context from Instagram' and its integration into platforms like Instagram Stories and WhatsApp suggests a push towards AI-generated content deeply embedded in social interactions. This could create novel content formats but also presents new challenges for content moderation and authenticity on these platforms.
Meta to release Muse Video for public use within 60 days
Meta has announced Muse Video alongside Muse Image, indicating it is a core part of their multimedia AI strategy. While currently in preview, the rapid rollout of Muse Image suggests a similar aggressive release schedule for Muse Video. It is plausible that Meta will make Muse Video publicly available within the next 60 days to compete in the generative video space.
Meta to integrate Muse Image into Advantage+ creative suite for ad automation within 30 days
Evidence suggests Meta's Muse Image is strategically positioned for ad automation, with a focus on its integration with the Advantage+ creative suite. Given the stated goal of reducing costs and effort for advertisers, it is highly probable that Meta will accelerate the integration of Muse Image into Advantage+ within the next 30 days to capitalize on this value proposition.
-
Meta launches Muse Ember, a cost-efficient AI model for agent tasks · 8 sources tracked
Meta has launched Muse Ember, its smallest and most efficient AI model, designed for high-volume, repetitive tasks performed by AI agents. The model is optimized for functions like tool calls, routing, and data extracti…
-
Meta AI Director Launches Ndoc to Replace Office, Workspace
Rohit Patel, a former director at Meta Superintelligence Labs, has left Meta after eight years to launch a new startup called Ndoc. Ndoc is designed as an AI-native platform intended to replace fragmented workflows asso…
-
Three open-weight AI models released with varying specs and licenses
A recent week saw the release of three open-weight AI models, each with distinct characteristics and licensing. MiniCPM-V 4.7, a 35B parameter model from OpenBMB, is notable for its ability to run on mobile devices like…
-
Meta releases 816B parameter Muse Spark Open model with custom license
Meta has released Muse Spark Open, an 816-billion-parameter mixture-of-experts model, with approximately 42 billion parameters active per token for efficiency. This model, designed for coding and agentic tasks with a 1M…
-
Meta's Muse Wonder AI models launch amid mixed reviews and trust concerns
Meta Superintelligence Labs has released its first frontier-class AI models, Muse Wonder and Muse Wonder Full, achieving competitive scores on various benchmarks, including the Artificial Analysis Intelligence Index whe…
-
Meta launches Muse Wonder frontier AI model family
Meta Superintelligence Labs (MSL) has launched its second-generation model family, Muse Wonder, approximately 14 months after the lab's inception. The Muse Wonder family includes two models: Muse Wonder, positioned as a…
-
Meta returns to AI frontier with Muse Wonder models
Meta has launched its new frontier-class AI models, Muse Wonder and Muse Wonder Full, marking a significant return to the forefront of AI development. These models demonstrate advanced capabilities in coding, long-conte…
-
Meta open sources Muse AI agent code for DIY hardware projects
Meta has released the open-source code for its AI agent, Muse, allowing hobbyists and developers to create their own AI hardware gadgets. The project encourages the use of off-the-shelf components like ESP32 boards and …
-
Meta's Muse AI leaks user address on Facebook Marketplace
Meta's AI agent, Muse, has reportedly leaked a YouTuber's home address to a stranger during a Facebook Marketplace transaction. The AI was given broad permissions to manage messages, and while the user provided his addr…
-
Meta's Muse Spark 1.3 launches with 1M context, dual pricing tiers
Meta has released Muse Spark 1.3, a closed multimodal reasoning model designed for long-running agentic tasks and coding. The model boasts a 1 million token context window and scores highly on benchmarks like DeepSWE v1…
-
Claude AI's email management capabilities introduce significant risks
Anthropic's Claude AI can now manage user inboxes, including sending and replying to emails without explicit approval. However, this functionality carries significant risks, such as prompt injection attacks where hidden…
-
Meta's Muse Spark 1.3 challenges top AI models with competitive pricing
Meta has released Muse Spark 1.3, an AI model designed for long-horizon agentic and coding tasks. The model shows improved efficiency, using fewer tool calls and tokens compared to its predecessor, Muse Spark 1.2. Muse …
-
Meta releases Muse Voice Transcribe, a real-time speech AI
Meta has launched Muse Voice Transcribe, a new real-time speech recognition model developed by Meta Superintelligence Labs. This model supports streaming audio transcription, identifies over 20 speakers, and detects the…
-
Meta unveils real-time AI transcription model for multi-speaker, multi-language audio
Meta has released a new AI transcription model capable of identifying multiple speakers and languages in real-time. This advanced model, developed by Meta Superintelligence Lab, aims to provide highly accurate transcrip…
-
Meta AI launches single real-time model for voice tasks
Meta AI has introduced Muse Voice Transcribe, a novel real-time audio perception model designed to handle speech recognition, speaker diarization, and endpointing within a single system. This model, which ranks highly o…
-
Meta's Muse Glimmer model now available on Fireworks AI platform
Fireworks AI has announced that Muse Glimmer, a new open-weight model from Meta Superintelligence Labs, is now available on their platform. This 30 billion parameter dense model is designed for continuous agent operatio…
-
Meta releases open-source agentic model Muse Glimmer for local use
Meta has released Muse Glimmer, an open-source agentic model designed for local execution on personal computers and Macs. This 30-billion parameter model, licensed under Apache 2.0, is optimized for "always-on" agent wo…
-
Together AI launches Meta's Muse Glimmer for agentic tasks · 3 sources tracked
Together AI has launched Muse Glimmer, an open-weight model developed by Meta Superintelligence Labs. This model is designed for complex, long-running agentic tasks, capable of reasoning, using tools, and recovering fro…
-
Meta releases Muse Glimmer, a 30B open-weight model for local AI agents
Meta has released Muse Glimmer, a 30-billion-parameter open-weight model optimized for local agentic workflows. This model is designed to run on consumer hardware, such as a single GPU, making it accessible for personal…
-
Alexandr Wang joins Meta Superintelligence Labs, advocates for small teams
Alexandr Wang, formerly of Scale AI, has joined Meta Superintelligence Labs and shared his perspective on team dynamics. He argues that smaller, agile teams are more effective than larger, bureaucratic ones. Wang's insi…