Meta ships Muse Wonder and Muse Wonder Full, scoring 52 and 56 on AA Intelligence Index
ArtificialAnlys · x · 2026-10-13
Meta released Muse Wonder and Muse Wonder Full, scoring 52 and 56 on the Artificial Analysis Intelligence Index v4.3, up from Muse Spark 1.3's 48 just six weeks earlier.
- Benchmarks: Wonder Full (max) improves on all ten evals, with the largest agentic gains: Terminal-Bench 4.0 from 33% to 57%, GDPval-AA v2.1 +176 Elo to 1857, AA-Briefcase +148 Elo. It places top-3 on six of ten evals, including #1 on CritPt (33%) and GDP.pdf (33%), with particular strength in agentic knowledge work and scientific reasoning.
- Efficiency: 54k output tokens per task, 10% fewer than its predecessor; Wonder Full uses 16% fewer input tokens than Wonder.
- Pricing: $4/$20 per million input/output tokens, $4.52 per task — 2.8x Muse Spark 1.3, cheaper than Claude models but pricier than Gemini 4 Argon and GPT-6.1 Sol.
- Wonder is available via Meta's first-party API; Wonder Full is limited to Muse Code users.
Related event: Meta Unveils Muse Wonder, Its Most Capable Frontier Model(92 posts)→
More from Models
- Allen AI's TutorMoments Benchmark Finds LLM Tutors Over-Help Students — allen_ai · 2026-10-13
- DeepSeek's flash endpoint may be quietly testing a new model, users speculate v4.1-pro — teortaxesTex · 2026-10-13
- 3D game demos hint at AI creativity that transfers to CAD, docs and UI tasks — Angaisb_ · 2026-10-13
- subquadratic's New Model Beats Jev in 6 of 7 Production Use Cases Before Launch — ryanallred · 2026-10-13
- Fresh ChatGPT accounts tell users they're in the 95th intelligence percentile, test finds — Polymarket · 2026-10-13
- Gemini 4 Argon still unavailable 11 days after announcement as Muse Wonder ships instantly — airesearch12 · 2026-10-13