{
  "data": null,
  "kind": "anchor.page",
  "links": {
    "api": "https://www.anchorterminal.com/api/v1/index.json",
    "html": "https://www.anchorterminal.com/people/alexandr-wang/",
    "json": "https://www.anchorterminal.com/people/alexandr-wang/index.json",
    "llms": "https://www.anchorterminal.com/llms.txt",
    "markdown": "https://www.anchorterminal.com/people/alexandr-wang/index.md",
    "mcp": "https://www.anchorterminal.com/mcp",
    "slim": "https://www.anchorterminal.com/people/alexandr-wang/index.min.md"
  },
  "markdown": "Chief AI Officer, Meta, leading Meta Superintelligence Labs.\n\nAlexandr Wang has been Meta's chief AI officer since 2025 and leads Meta Superintelligence Labs. He co-founded Scale AI, a data labelling and model evaluation company, in 2016 and was its chief executive.\n\n## What Alexandr builds\n\n- [Llama Guard 4](https://www.anchorterminal.com/tools/llama-guard.md), graded D, 49.5/100, in Guardrails \u0026 safety filters. Llama Guard 4 is Meta's 12-billion-parameter open-weight safety classifier for text and images. It labels a prompt or a model response safe or unsafe against 14 hazard categories, and the owner runs it on a GPU. Chief AI Officer, Meta, leading Meta Superintelligence Labs, researched 2026-10-28.\n- [LlamaFirewall](https://www.anchorterminal.com/tools/llamafirewall.md), graded D, 51.1/100, in Guardrails \u0026 safety filters. LlamaFirewall is Meta's open-source Python library for screening an AI agent's inputs, tool results and outputs. It runs scanners for prompt injection, hidden characters, insecure generated code and goal drift, and returns allow, block or human review. Chief AI Officer, Meta, leading Meta Superintelligence Labs, researched 2026-10-28.\n- [Meta Model API](https://www.anchorterminal.com/tools/meta-model-api.md), graded D, 52.5/100, in Model APIs \u0026 inference. Meta Model API is Meta's pay-as-you-go API for its hosted Muse models, including Muse Wonder and Muse Spark, through OpenAI-style Responses and Chat Completions endpoints and an Anthropic-style Messages endpoint, plus image, speech and segmentation routes. Chief AI Officer, Meta, leading Meta Superintelligence Labs, researched 2026-10-28.\n- [Muse Code](https://www.anchorterminal.com/tools/muse-code.md), graded C, 61.4/100, in Agent harnesses. Muse Code is Meta's coding agent for the terminal and CI, built for its Muse models. It plans, edits files and runs commands under approvals and an OS sandbox, interactively, headless through `muse exec` or from an SDK. Chief AI Officer, Meta, leading Meta Superintelligence Labs, researched 2026-10-28.\n- [Muse Glimmer 30B](https://www.anchorterminal.com/tools/muse-glimmer-30b.md), graded BB, 70/100, in Open-weight models. Muse Glimmer 30B is Meta's open-weight dense multimodal model for local agent work, 29.6B parameters with text and image input and a 128K-token context, published on Hugging Face under Apache 2.0 with GGUF and ExecuTorch builds. Chief AI Officer, Meta, leading Meta Superintelligence Labs, researched 2026-10-28.\n- [Muse Spark Open](https://www.anchorterminal.com/tools/muse-spark-open.md), graded D, 53.5/100, in Model APIs \u0026 inference. Muse Spark Open is Meta's open-weight multimodal model for coding and agentic work, with 816B total and 42B active parameters and a 1M-token context, released on 15 October 2026 under the Muse Model Licence for self-hosting. Chief AI Officer, Meta, leading Meta Superintelligence Labs, researched 2026-10-28.\n\n## Works with\n\n- [Abraham Montilla](https://www.anchorterminal.com/people/abraham-montilla/), Llama Guard 4 release/model-card contributor (GitHub amontilla0)\n- [Andrew Bosworth](https://www.anchorterminal.com/people/andrew-bosworth/), CTO, Meta, and head of Reality Labs\n- [Andrew McCollum](https://www.anchorterminal.com/people/andrew-mccollum/), Co-founder, Facebook, worked there until September 2007\n- [Chris Hughes](https://www.anchorterminal.com/people/chris-hughes/), Co-founder and spokesman, Facebook, left in 2007\n- [Connor Treacy](https://www.anchorterminal.com/people/connor-treacy/), Meta Product Manager; Meta Model API launch-guide co-author\n- [Dustin Moskovitz](https://www.anchorterminal.com/people/dustin-moskovitz-2/), Co-founder, Facebook, left in 2008\n- [Eduardo Saverin](https://www.anchorterminal.com/people/eduardo-saverin/), Co-founder and first CFO, Facebook, no longer with the company\n- [Josh Walters](https://www.anchorterminal.com/people/josh-walters/), Developer marketing; Meta Model API launch-guide co-author\n- [Joshua Saxe](https://www.anchorterminal.com/people/joshua-saxe/), LlamaFirewall research coauthor; former Meta security AI leader\n- [Mahesh Pasupuleti](https://www.anchorterminal.com/people/mahesh-pasupuleti/), Senior Engineering Leader, Meta Superintelligence Labs, AI Safety and Alignment; led Llama Guard development\n- [Mark Zuckerberg](https://www.anchorterminal.com/people/mark-zuckerberg/), Founder, chairman and CEO, Meta\n- [Sahana Chennabasappa](https://www.anchorterminal.com/people/sahana-chennabasappa/), AI Security Engineer; LlamaFirewall paper first-listed author\n- [Tatiana Shavrina](https://www.anchorterminal.com/people/tatiana-shavrina/), AI/ML Research Lead; Muse Glimmer research team\n\n## Questions\n\n### Who is Alexandr Wang?\n\nAlexandr Wang has been Meta's chief AI officer since 2025 and leads Meta Superintelligence Labs. He co-founded Scale AI, a data labelling and model evaluation company, in 2016 and was its chief executive.\n\n### What does Meta make?\n\nAnchor Terminal lists Llama Guard 4, LlamaFirewall, Meta Model API, Muse Code, Muse Glimmer 30B and Muse Spark Open from Meta. Llama Guard 4 is Meta's 12-billion-parameter open-weight safety classifier for text and images. It labels a prompt or a model response safe or unsafe against 14 hazard categories, and the owner runs it on a GPU.\n\n### How are Alexandr's tools graded?\n\nLlama Guard 4 is graded D, 49.5 out of 100; LlamaFirewall is graded D, 51.1 out of 100; Meta Model API is graded D, 52.5 out of 100; Muse Code is graded C, 61.4 out of 100; Muse Glimmer 30B is graded BB, 70 out of 100; Muse Spark Open is graded D, 53.5 out of 100, on Anchor Terminal's benchmark, which scores every listing against the same checklist.\n\n## Sources\n\n- https://www.google.com/search?q=%22Llama+Guard%22+%22Meta%22+%22led+development%22\n- https://www.linkedin.com/in/maheshp9\n- https://github.com/meta-llama/PurpleLlama/blob/main/Llama-Guard4/12B/MODEL_CARD.md\n- https://github.com/meta-llama/PurpleLlama/commit/cd9fe65792343c285730c967f888a7373d7e6c17\n- https://github.com/amontilla0\n- https://www.meta.com/en-gb/about/leadership/mark-zuckerberg/\n- https://dev.meta.ai/\n- https://www.google.com/search?q=Meta+company+LinkedIn\n- https://dev.meta.ai/llama/docs/community-support-and-resources/\n- https://github.com/meta-llama/PurpleLlama\n- https://dev.meta.ai/llama/docs/model-cards-and-prompt-formats/llama-guard-4/\n- https://ai.meta.com/research/publications/llamafirewall-an-open-source-guardrail-system-for-building-secure-ai-agents/\n- https://www.linkedin.com/in/sahana-chennabasappa-69a77576\n- https://x.com/joshua_saxe\n- https://github.com/meta-llama/PurpleLlama/blob/main/CONTRIBUTING.md\n- https://github.com/meta-llama/PurpleLlama/tree/main/LlamaFirewall\n- https://en.wikipedia.org/wiki/Meta_Platforms\n- https://en.wikipedia.org/wiki/Mark_Zuckerberg\n- https://dev.meta.ai/resources/blog/build-with-muse-spark\n- https://uk.linkedin.com/in/connort\n- https://www.linkedin.com/posts/joshmwalters_llama-meta-genai-activity-7358885091185426434-VPdf\n- https://www.linkedin.com/in/joshmwalters\n- https://dev.meta.ai/support/\n- https://dev.meta.ai/docs/muse-glimmer\n- https://uk.linkedin.com/in/tatiana-shavrina/en\n- https://dev.meta.ai/help\n- https://dev.meta.ai/support\n\nThese are the people building the tools we list, named for their public role in the company or project behind a listing, from public pages, with the date it was researched. They don't work for Anchor Terminal (our own team is on the about page, /about/#author). We show public roles, public profiles and what the listing pages show, nothing private. To correct something, send an about or a photo, or have a page removed, write to privacy@anchorterminal.com.\n",
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit Anchor Terminal and link the page or API address you took it from (https://www.anchorterminal.com/terms/#data)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-10-28",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.7",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  },
  "page": {
    "breadcrumbs": [
      {
        "name": "Home",
        "url": "https://www.anchorterminal.com/"
      },
      {
        "name": "People building AI tools",
        "url": "https://www.anchorterminal.com/people/"
      },
      {
        "name": "Alexandr Wang",
        "url": ""
      }
    ],
    "description": "Alexandr Wang is chief AI Officer, Meta at leading Meta Superintelligence Labs. Llama Guard 4, LlamaFirewall, Meta Model API, Muse Code, Muse Glimmer 30B and Muse Spark Open on Anchor Terminal, with grades, an about and public profiles.",
    "facts": null,
    "h1": "Alexandr Wang",
    "image": "https://www.anchorterminal.com/assets/og/default.png",
    "path": "/people/alexandr-wang/",
    "published": "",
    "section": "people",
    "title": "Alexandr Wang, Meta | Anchor Terminal",
    "toc": null,
    "updated": "2026-10-28",
    "url": "https://www.anchorterminal.com/people/alexandr-wang/"
  },
  "tokens": {
    "markdown": 2100,
    "slim": 1830
  },
  "version": 1
}