Meta Open-Sources Muse Glimmer, a 30B Local Agentic Model

TLDR

Meta on August 10, 2026 shipped Muse Glimmer, a 30-billion-parameter dense model from Meta Superintelligence Labs, with weights released under Apache 2.0 on Hugging Face. The model is built for always-on local agent workflows—function calling, local coding, multi-step tool use, and LLM-as-a-judge—on a Mac or PC with a single consumer GPU (quantized under 20 GB with headroom for KV cache and perception). Same day, CEO Mark Zuckerberg published a **14-page** essay (“The Future is for Everyone”) urging U.S. policy to ease open-weight friction versus Chinese labs and criticizing extreme concentration of AI power—without naming OpenAI and Anthropic.

What shipped

Item	Detail (Meta research blog / HF / AA)
Model	Muse Glimmer — 30B dense; multimodal text + images via perception encoder
License	Apache 2.0 (Meta’s most permissive open release yet vs prior Llama License)
Context	128K token context (plus extension)
Target	Local agents: long-horizon tasks, precise tool schemas, failure recovery / retry
Memory path	Full precision >55 GB; ~4-bit quant brings LM under ~20 GB for 24–32 GB consumer envelopes
Speed stack	Speculative decoding with DFlash-style drafter; validated on M4/M5 Max Macs and RTX 5090
Training	Logit distillation from Muse Spark; mid-train agent traces; SFT + on-policy distillation + RL
Access	Weights: meta-models/Muse-Glimmer-30B on Hugging Face; Ollama / LM Studio / llama.cpp / MLX / ExecuTorch integrations landing; serving via vLLM, SGLang, Together, Fireworks, OpenRouter
Safety	Evaluated under Meta’s Advanced AI Scaling Framework for open-weight release

Zuckerberg’s same-day messaging also previewed opening weights for Muse Spark 1.2 (Meta’s closed frontier coding/chat tier) and “even bigger models coming soon,” plus a $1 billion community fund tied to data-center build-out—distinct product lines from the Glimmer ship itself.

Ranking context

Lab-reported: Meta positions Glimmer as strong vs Gemma 4 31B and Qwen3.6-27B on agentic suites (DeepSearch QA, MCP-Atlas, τ-Bench, SWE-Bench) with methodology notes in the launch report—prefer the linked methodology over raw marketing charts.

Independent (near launch):

• Artificial Analysis (Aug 10 analysis): Muse Glimmer scores 35 on the Intelligence Index—21 points above Llama 4 Maverick (14), Meta’s prior open-weight line; ~5 points above Gemma 4 31B (Reasoning) in AA’s size-class comparison. Openness Index ~44 (Apache 2.0 + methodology disclosure), matched with peers such as DeepSeek V4 Flash / GLM-5.2 class open models.
• LMArena / Arena: live boards may lag a same-day open drop; do not invent Elo—check muse-glimmer listings as votes accumulate.
• Not a frontier closed-model challenger on AA’s top band (Opus / Sol class); the product thesis is local agent throughput + open weights, not cloud frontier chat Elo.

Product-line placement

Distinct from Muse Code / Spark 1.2 (Aug 5 closed coding stack), Muse Image, and the Aug 5 cyber-eval breach story. Glimmer is Meta’s first open-weights MSL release since Llama 4—a deliberate return to the open lane after a closed-first Superintelligence Labs period.

Why this story matters

Open-weight competition is no longer a China-only story. Apache 2.0 at 30B for offline agents answers cost, cyber-refusal, and sovereignty concerns that closed APIs create—while Zuckerberg’s essay frames open weights as U.S. industrial policy, not only developer goodwill. Watch: whether Spark 1.2 weights actually ship; Arena and AA movement after community quant/fine-tunes; and whether enterprises adopt Glimmer as the non-Chinese open default.

Sources

• Meta AI Research: Introducing Muse Glimmer (August 10, 2026)
• Hugging Face: meta-models/Muse-Glimmer-30B
• Artificial Analysis: Muse Glimmer benchmarks (August 10, 2026)
• Reuters: Meta launches new AI model as Zuckerberg champions open-weight push (August 10, 2026)
• CNBC: Meta Muse Glimmer open-weight AI (August 10, 2026)
• Times of AI prior: Muse Code and Muse Spark 1.2 (August 5, 2026)