Meta's Small Model Muse Spark Outranks Astra DeepSwe as Unreleased Model Climbs Benchmarks
altryne · x · 2026-09-04
Blogger altryne notes that while OpenAI may have won the day, Meta — not Anthropic, Google, or xAI — took second place. Meta's Muse Spark scored above Astra DeepSwe, and an unreleased Meta model is already showing higher on Artificial Analysis (though confusingly). Notably, Spark is Meta's small model, hinting at bigger things from its unreleased flagship. An interesting shuffle at the top of AI, per the author.
More from Models
- Martian says routing across 44 LLMs cuts errors 46% vs best single model on 16 benchmarks — Arindam_1729 · 2026-09-04
- Ex-OpenAI safety lead Miles Brundage: if your primary emotion on AI isn't concern, you're misreading it — Miles_Brundage · 2026-09-04
- Gary Marcus on GPT-6 Astra: symbolic world models are vindication, but no proof of AGI — GaryMarcus · 2026-09-04
- Ex-Google X Quant Guillaume Verdon Says OpenAI Is 'Kinda Back' — But Vibes, Not Benchmarks, Will Decide — beffjezos · 2026-09-04
- Mystery model "Astra" reportedly beats 5.6 Sol Pro on FrontierMath T4 — ctjlewis · 2026-09-04
- Inkling and Inkling-Small models get a free inference tier, and it's staying — simonguozirui · 2026-09-04