Meta launches Muse Wonder models and Muse Code coding agent, priced $5-$200/month
jyangballin · x · 2026-10-13
Meta released its largest models to date, Muse Wonder and Muse Wonder Full, claiming frontier performance in personal/professional agents and coding, alongside the Muse Code coding agent (macOS/Windows, installed via curl install.sh | bash).
Pricing: $5/mo (basic), $15/mo (5x usage), $50/mo (20x), $200/mo (80x).
Tech: Muse Code is multi-agent by default (parallel workers + background reviewers), fully auditable, and co-trained with Muse Spark; the API offers a 1M context window.
Benchmarks (vs Claude Opus 5.5 / GPT-6 Astra / Claude Fable 5.1):
- SWE-Marathon: 57.5%, well ahead of the runner-up at 53.1%
- DeepSWE: 79.0%, first place
- Personal Superintelligence Index: 59.1%, top
- Terminal-Bench 4.0-AA: 57.1%, slightly below Claude Opus 5.5 (59.6%) and GPT-6 Astra (59.1%)
- Leads MemoryBench (85.5) and JobBench (75.3)
Meta also announced a global AI developer hackathon with $1M in prizes.
Related event: Meta Unveils Muse Wonder, Its Most Capable Agentic Frontier Model(88 posts)→
More from coding & agent
- PageIndex SDK v0.4 claims tree indexing now cheaper than vector indexing by ~15% — solyarisoftware · 2026-10-13
- Economist: agentic coding lets you write your own Fortran optimizers, not just ask for Stata packages — Apoorva__Lal · 2026-10-13
- Tencent open-sources Agent Memory: local long-term memory, 61% fewer tokens, 48%→76% accuracy — solyarisoftware · 2026-10-13
- Testing AI agents with simulations: replay real traces, 50 curated sims beat 1000 random ones — HamelHusain · 2026-10-13
- How humans and Copilot agents split the work to build a GitHub Copilot beginner course — DanWahlin · 2026-10-13
- Engineers distill their AGI debate into an essay on self-verifying, continuously improving agents — dotey · 2026-10-13