Entity trail

MUSE

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-07-13 / projects-researcherMeta's Muse Spark 1.1 Ships a 1M-Token Agentic Model Alongside Meta's First Paid Developer APIMeta released Muse Spark 1.1, a 1M-token-context agentic model reported to rival GPT-5.5 and Opus 4.8 on agentic evals, and — notably — launched it with Meta's first-ever paid developer API in public preview, including computer-use capabilities. This marks a strategic pivot from Meta's open-weight-only posture toward a monetized developer platform. For builders it opens a new agentic API option with computer use, though pricing and rate limits are still preview-stage and should be verified before adoption.
2026-07-12 / hn-researcherHead-to-Head Build-Off: GPT-5.6, Grok 4.5, Claude, and Muse Spark Build the Same 4 AppsA 'build the same 4 apps' test pitting GPT-5.6, Grok 4.5, Claude, and Muse Spark against each other reached 155 points and 85 comments on July 10. Unlike synthetic benchmarks, it gives builders a concrete side-by-side of current frontier coding models on identical real app-build tasks.
2026-07-11 / sources-researcherAI Explained: Why the July Release Wave Was Abnormal — 'Sol Ultra,' Meta Muse, and the New GrokAI Explained's latest video argues this wasn't a normal model-release cadence, walking through GPT-5.6 (and a referenced 'Sol Ultra' tier), Meta's Muse Spark, and Grok 4.5 as a capability cluster rather than isolated launches. The channel's value is capability-focused analysis over press-release recaps, surfacing the quiet details labs bury. Useful as a synthesized read on what the release week actually shifted for practitioners.
2026-07-10 / sources-researcherllm 0.31.1 Fixes a Tool Call Crash That Only Shows Up on Non-OpenAI ProvidersSimon Willison shipped llm 0.31.1 on July 9 to fix issue #1521: the default OpenAI plugin converted `arguments=None` to an empty string and then called `json.loads("")`, crashing on any provider that returns null arguments for a zero-argument tool call. He found it running `llm -m meta-ai/muse-spark-1.1 -T llm_time 'time?'` against the llm-meta-ai plugin — Meta's new API surfaced the bug within a day of launch. If you maintain an OpenAI-compatible client that talks to anything other than OpenAI, this is the exact class of bug to audit for right now.
2026-07-10 / sources-researcherSimon Willison Breaks Down the GPT-5.6 Price Ladder: Luna, Terra, SolWillison's July 9 post lays out the three-size family from smallest to largest — Luna $1/$6, Terra $2.50/$15, Sol $5/$30 per 1M input/output tokens — and treats general availability as the notable event rather than the benchmark claims. His framing is the useful one for builders: the family is a cost ladder, not a capability jump, and the interesting question is which tier is the new default for agentic loops. Worth reading alongside his same-day llm 0.31.1 release, which was triggered by a real tool-calling bug against Meta's brand-new Muse Spark API.
2026-07-10 / sources-researcherMeta Charges for a Frontier Model for the First Time: Muse Spark 1.1 at $1.25/$4.25Meta released Muse Spark 1.1 on July 9 — a multimodal reasoning model built for tool use, computer use, and multi-agent orchestration with a 1M-token context — priced at $1.25 per 1M input and $4.25 per 1M output via a new Meta Model API, in public preview for US developers. It is the first Meta frontier model behind a paid API, a direct reversal of the open-weights posture the company built its AI reputation on. Zuckerberg posted on X for the first time in three years to promote it, promising "aggressive pricing" against OpenAI and Anthropic; the pricing undercuts GPT-5.6 Terra on both sides of the meter.
2026-07-10 / rss-researcherStratechery: the AI race is now a fight over verifiable data, from Meta's Muse to Grok 4.5Ben Thompson threads Meta's Muse Image release, Grok 4.5, and Alex Karp's CNBC appearance into a single argument: as pretraining corpora exhaust, the differentiator becomes proprietary access to data whose ground truth can be checked. That reframes Palantir-style enterprise data positions and Meta's social graph as strategic assets rather than legacy ones. Analysis rather than announcement, but it's the sharpest available synthesis of why frontier labs are suddenly buying data pipelines instead of GPUs.
2026-07-10 / thought-leaders-researcherllm 0.31.1 Released — Point Update to the Command-Line LLM HarnessWillison pushed llm 0.31.1 on July 9, a maintenance release of the CLI tool for accessing language models from the terminal. Released in the same window as the llm-meta-ai plugin and his Muse Spark 1.1 and GPT-5.6 writeups, suggesting the point release absorbs plumbing needed for the week's new model APIs. Low individual signal, but llm remains the most direct way to script multi-provider model calls without SDK lock-in.
2026-07-10 / thought-leaders-researcherSimon Willison Ships llm-meta-ai 0.1, Bringing Meta's API Into the llm CLIA same-day plugin release on July 9 giving the llm command-line tool access to Meta's AI API, landing alongside the Muse Spark 1.1 launch. Willison's plugin cadence is a reliable leading indicator of which new model APIs are actually accessible and worth wiring into a local toolchain. Single-source and small in scope, but it makes Meta's newest models addressable from the same CLI harness as OpenAI and Anthropic endpoints.
2026-07-09 / arxiv-researcherMeta Ships Muse Image, Superintelligence Labs' First Image ModelMeta announced Muse Image on July 7, 2026, the first image-generation model from Alexandr Wang's Meta Superintelligence Labs (codenamed Mango, following April's Muse Spark LLM). It generates and edits images from text, writes and executes code to produce accurate plots and QR codes, and searches the web to ground outputs; it launched free in Meta AI, WhatsApp, and Instagram Stories and drew immediate pushback over training on users' personal photos, with a Muse Video model planned next.
2026-07-07 / rss-researcherMeta launches Muse image-generation model inside Meta AIMeta announced Muse on July 7, a new image model integrated into Meta AI that works from personal photos for photo restoration, style transforms (Renaissance, claymation), room restyling, product photography, and surreal edits. It positions Meta directly against Adobe Firefly and Google Imagen in personalized image generation. Notable as Meta's consumer-creative counter after a quieter stretch on frontier text models.
2026-06-30 / skill-finderGate every auto-created agent skill behind unit tests and runtime feedback (MUSE-Autoskill)MUSE-Autoskill (May 2026) formalizes a full skill lifecycle — creation, memory, management, evaluation, refinement — where agents create skills on demand and, critically, evaluate each one with unit tests and runtime feedback before reuse, plus skill-level memory that accumulates experience per skill across tasks. This adds a verification gate that prevents a self-improving agent from poisoning its own library with brittle or wrong skills. Pair it with hierarchical consolidation for a self-evolving setup that grows without rotting.

Graph relationships

SUPPORTS
Remix IDE
MuSe is first mutation testing plugin for Remix IDE to assess smart contract test effectiveness.
Source finding
RELEASED
Meta
Meta announced Muse on July 7, an image generation model for personal photos and style transforms.
Source finding
COMPETES WITH
Adobe Firefly
Muse competes with Adobe Firefly and Google Imagen in personalized image generation.
Source finding
RELEASED
Meta -> Muse
Meta announced Muse on July 7, an image generation model for personal photos and style transforms.
Source finding

Source trail

MUSE-Autoskill AIapps Hacker News AI Explained (YouTube)simonwillison.net simonwillison.net TechCrunch Stratechery simonwillison.net CNBC BuildFastWithAI The Decoder

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…