Meta AI

18 releases covered on ThursdAI · ai.meta.com ↗

July 2026

Meta AI
New Models

Muse Spark 1.1 & Meta Model API

Meta launches Muse Spark 1.1 and its first paid Meta Model API

Mark Zuckerberg returned to X (35 seconds into the ThursdAI live show) to announce Muse Spark 1.1: a 1M-token-context agentic model that rivals GPT-5.5 and Opus 4.8 on agentic evals, claiming #1 on MCP Atlas, JobBench, Humanity's Last Exam and Finance Agent V2. It ships with Meta's first-ever paid developer API in public preview ($20 free credits, US-only at launch), computer use across desktop, browser and mobile, and parallel subagent delegation. On the held-back Vals AI Harvey legal-agent benchmark it scores 20% against Fable's 11%. Replit, Cline and Box are early partners. No open weights.

$1.25/$4.25 Per 1M tokens (in/out)1M Token context window20% vs 11% Harvey Legal Agent Bench vs Fable
Alexandr Wang announcement ↗Meta blog ↗AI at Meta ↗
🎙️ Hear our coverage →
#frontier-models#agents#api
Meta AI
New Models

Muse Image & Muse Video

Meta Superintelligence Labs ships Muse Image and previews Muse Video

MSL's first media-generation models: Muse Image is live in the Meta AI app, Instagram Stories (US) and WhatsApp, with agentic generation that calls web search and code execution, multi-reference composition, and Instagram social-context conditioning. Muse Video shares the same pretraining base and adds native audio, debuting at #3 on Arena text-to-video while Muse Image lands #2 on image. There is no public API, and public Instagram accounts are opted in to @-mention remixing by default.

#2 Arena text-to-image debut#3 Arena text-to-video debut1280 Arena image score
X announcement ↗Blog ↗
🎙️ Hear our coverage →
#image-gen#video-gen#consumer-ai

May 2026

Meta AI
Major Features & Updates

Muse Spark voice conversations

Meta launches Muse Spark voice conversations across its apps and glasses

Meta rolled out Muse Spark-powered voice conversations across the Meta AI app, WhatsApp, Instagram, Facebook, and Ray-Ban Meta glasses. The feature includes real-time image generation, live camera AI, and instant Reels/maps integration. Alex tested it live and called it surprisingly good, the first big consumer ship from Meta Superintelligence Labs.

X announcement ↗Announcement ↗
🎙️ Hear our coverage →
#voice-ai#consumer-ai#multimodal

April 2026

Meta (Meta Superintelligence Labs)
New Models

Muse Spark

Meta launches Muse Spark, first model from Meta Superintelligence Labs

Meta dropped Muse Spark mid-show, the debut model from Meta Superintelligence Labs. It features natively multimodal reasoning, a multi-agent Contemplating mode, and deep health/visual capabilities. Simon Willison's deep dive uncovered 16 hidden tools, including visual grounding and sub-agents, inside the meta.ai chat UI.

AI at Meta announcement on X ↗Introducing Muse Spark (Meta blog) ↗MSL announcement ↗Simon Willison's deep dive on the 16 hidden tools ↗
🎙️ Hear our coverage →
#frontier-models#multimodal#agents

December 2025

November 2025

Meta AI
New ModelsOpen weights

SAM 3

Meta SAM 3: open-vocabulary segmentation and tracking in video

Meta's Segment Anything Model 3 adds open-vocabulary segmentation with text and exemplar prompts, letting you click or type to segment and track any object across images and video. The panel demoed it live on golden retriever videos, and it ships openly as part of Meta's open-source push.

🎙️ Hear our coverage →
#vision#open-source
Meta AI
New ModelsOpen weights

SAM 3D

SAM 3D turns single photos into 3D objects and human bodies

Released alongside SAM 3, SAM 3D reconstructs 3D objects and full human bodies from a single image with surprisingly high quality. It extends the Segment Anything family from 2D segmentation into single-image 3D reconstruction.

🎙️ Hear our coverage →
#vision#world-models#open-source
Meta AI
New ModelsOpen weights

Omnilingual ASR

Meta releases Omnilingual ASR covering 1,600+ languages

Meta released Omnilingual ASR, an Apache 2.0 speech recognition family supporting over 1,600 languages, including 500+ never before served by any ASR system, with character error rate under 10% for 78 languages. The release includes an open corpus of 500k+ rows of transcribed audio, and the 1B model was praised as a near drop-in state-of-the-art replacement on Hugging Face.

1600+ Languages Supported
AI at Meta announcement on X ↗Meta blog post ↗Research paper ↗Omnilingual ASR corpus on Hugging Face ↗
🎙️ Hear our coverage →
#voice-ai#open-source

October 2025

Meta AI (PyTorch)
Dev ToolsOpen weights

TorchForge

TorchForge: PyTorch-native library for scalable RL post-training

Meta's PyTorch team, in collaboration with Weights & Biases/CoreWeave and Stanford, introduced TorchForge, a PyTorch-native library for scalable reinforcement-learning post-training and agent development. Built for massive GPU runs (W&B/CoreWeave provided 520 H100s) and competing with Ray via tools like the Monarch scheduler.

520 H100s provided for development runs
Blog ↗
🎙️ Hear our coverage →
#training#coding

September 2025

Meta AI
New ModelsOpen weights

Code World Model (CWM)

Meta releases 32B Code World Model for agentic code reasoning

Meta released CWM, a 32B open-weights research model trained to internally model code execution, aimed at agentic code reasoning rather than plain code completion. The weights are on Hugging Face under facebook/cwm, giving the open-source community a new approach to code world modeling.

X ↗HF ↗
🎙️ Hear our coverage →
#open-source#coding#agents
Meta AI & Hugging Face
Benchmarks & EvalsOpen weights

Gaia2 + ARE

Gaia2 agent benchmark and Agents Research Environments released

Meta and Hugging Face released Gaia2, a follow-up agent benchmark, together with ARE (Agents Research Environments) for testing agents in dynamic, asynchronous settings. It fed the episode's recurring concern that evaluation has to keep up whenever agent product claims get ambitious.

X ↗HF ↗
🎙️ Hear our coverage →
#benchmarks#agents
Meta AI
Products & Apps

Meta AI Glasses with Display

Meta Connect: new AI glasses with a display and neural control interface

At Meta Connect, Meta unveiled new AI glasses featuring a built-in display, a neural wristband control interface, and a new AI mode. The panel treats the glasses as an interface milestone, arguing the product surface for AI is shifting from apps to display-equipped wearables.

X Recap ↗
🎙️ Hear our coverage →
#infrastructure#consumer-ai#multimodal

July 2025

Meta
Also Released

Meta Superintelligence Labs (MSL)

Meta launches Superintelligence Labs with up to $300M comp packages

Zuckerberg formally assembled Meta Superintelligence Labs, recruiting a dream team of researchers from OpenAI and other labs with rumored compensation packages of up to $300M. The panel treated the spree as proof that the AI talent war has entered full wartime economics, debating whether money alone can buy research momentum.

$300M Rumored Meta packages
MSL hires tracker (spreadsheet) ↗
🎙️ Hear our coverage →
#industry

May 2025

April 2025

Meta AI
New ModelsOpen weights

Llama 4 (Scout & Maverick)

Meta drops Llama 4 Scout (109B) and Maverick (400B) open-weights MoE models

Meta released the long-awaited Llama 4 family in a chaotic Saturday drop: Scout (17B active / ~109B total, 16 experts) and Maverick (17B active / ~400B total, 128 experts), with a 2T-parameter Behemoth still in training. The models are multimodal, multilingual MoE architectures trained on ~30T tokens with FP8 and interleaved attention (iRoPE), claiming 10M context for Scout and 1M for Maverick. The release was marred by drama: the LMArena version differed from the released model, and the community criticized the lack of small local-friendly sizes.

10M Stated context window for Llama 4 Scout288B Active parameters of unreleased Behemoth (2T total)17B Active parameters for both Scout and Maverick
Meta blog: Llama 4 multimodal intelligence ↗Hugging Face: meta-llama ↗Try it at meta.ai ↗
🎙️ Hear our coverage →
#open-source#architecture#multimodal
Meta AI
Papers & Research

MoCha

Meta's MoCha generates movie-grade talking AI characters from speech and text

Meta GenAI researchers published MoCha, a model that generates stunningly realistic, movie-grade talking characters directly from speech plus text. Co-author Cong Wei joined the show to discuss the work, which points at AI actors entering Hollywood-quality territory.

MoCha project page ↗
🎙️ Hear our coverage →
#video-gen#research