Meta’s Muse Wonder just dropped and it’s cooking on agent work SWE-Marathon: 57.5 vs 46.3 opus automationbench: 45.4 vs 41.6 GPT-6 Astra MemoryBench: 85.5 vs 81.5 opus deepSWE: 79.0 vs 74.2 AA index still has opus on top (58 vs 56) but the long-horizon numbers are real meta is back in the fight

Meta’s Muse Wonder just dropped and it’s cooking on agent work SWE-Marathon: 57.5 vs 46.3 opus automationbench: 45.4 vs 41.6 GPT-6 Astra MemoryBench: 85.5 vs 81.5 opus deepSWE: 79.0 vs 74.2 AA index still has opus on top (58 vs 56) but the long-horizon numbers are real meta is back in the fight

🌐 链接: https://x.com/nex_ify/status/2109857788214046991
🏷️ 分组: X｜为你推荐
🕒 时间: 2026-10-13 12:04:30

X (formerly Twitter)

Nexx (@nex_ify) on X

Meta’s Muse Wonder just dropped and it’s cooking on agent work

SWE-Marathon: 57.5 vs 46.3 opus
automationbench: 45.4 vs 41.6 GPT-6 Astra
MemoryBench: 85.5 vs 81.5 opus
deepSWE: 79.0 vs 74.2…