gemini-4IMAGE
Gemini 4 Argon 对比 GPT-6 Astra 与 Fable 5.1 跑分
图表显示 Argon 在 DeepSWE 77.9%、Vibe Code Bench 91.9%、LABBench 2 88.8%、GraphWalks 84.2% 领先,但 Astra 仍在 FrontierSWE 等占优。
Priya
@Priyannkaaaa28 天前收录原推发布于 2026.10.01

Priya @Priyannkaaaa原推发布于 2026.10.01
💀💀Google new model is totally wild Gemini 4 Argon vs GPT-6 Astra vs Claude Fable 5.1 👀 Google’s new Gemini 4 Argon is putting up some crazy numbers. DeepSWE v1.1: 77.9% Argon > Astra 74.1% > Fable 67.4% Vibe Code Bench: 91.9% Argon > Fable 90.3% > Astra 89.6% LABBench 2: 88.8% Argon > Astra 85.4% > Fable 68.6% GraphWalks (256K–1M): 84.2% Argon > Astra 71.8% > Fable 65.0% But Astra still leads on FrontierSWE, Terminal-Bench and OSWorld-2.0. Gemini 4 Argon just made the frontier even more competitive. 🔥
x.com/priyannkaaaa/status/2105393103511372120采集于 2026.10.01 10:01