Meta's Muse Spark 1.3 hits 61-62 on AA Intelligence Index, cheapest per-task at its level
ArtificialAnlys · x · 2026-09-03
Meta released Muse Spark 1.3, its fourth release in five months. The xhigh variant scores 61 on the Artificial Analysis Intelligence Index (up from 57), tying GPT-5.6 Sol (max) and Grok 4.6 (high); the limited-preview max variant scores 62, behind only Claude Fable 5.1 and Claude Opus 5.
- Gains are driven by agentic work: Tau3-Bench Banking 35%→47%/52% (#1), Terminal-Bench 2.1 at 85-86%, GDPval-AA v2 Elo +94/+139.
- xhigh costs $0.55 per Intelligence Index task at unchanged $1.25/$4.25 per 1M token pricing — the cheapest among models scoring 59+, though up from 1.2's $0.40 due to 57% more input tokens.
- Scientific reasoning also rose: CritPt +8 to 26%, GPQA Diamond +4 to 94%.
More from Models
- GPT-6 'Astra' expected tomorrow: codename spotted in API, rumored 4x GPT-5.6 Sol — daniel_mac8 · 2026-09-03
- Greg Brockman appears to confirm o1, o3 and GPT-5 all built on the GPT-4o base model — flowersslop · 2026-09-03
- Heavy user says GPT-5.3's instruction following is OpenAI's special sauce over Claude — brandon_galang · 2026-09-03
- Early user test: Muse Spark 1.3 oneshots a full website frontend — alexandr_wang · 2026-09-03
- Researcher: Looped Transformers add no real recurrence, just tied-parameter depth — savvyRL · 2026-09-03
- Muse Spark 1.3 lands on OpenCode Zen for free, CEO urges devs to try it — alexandr_wang · 2026-09-03