gemini-4IMAGE

Gemini 4 Argon scores 53 on AA index, matching Astra

A Chinese post rounds up Argon's results: an AA index of 53 matching GPT-6 Astra, a new DeepSWE SOTA at 77.9%, a 1M-token output limit and a per-task cost at 60% of Astra's.

SuSu_酥酥👅

@NFT_ChenArchived 28 days agoOriginally posted 2026.10.01

Gemini 4 Argon scores 53 on AA index, matching Astra

SuSu_酥酥👅 @NFT_ChenOriginally posted 2026.10.01

Open original on X

🫡全体起立！谷歌Gemini 4 Argon今晚把锅给砸穿了！ Gemini 4 Argon AA智能指数：53分，追平GPT-6 Astra，卡在帕累托最优线上！对比当前前沿： 🔹Claude Opus 5.5（58）/ Sonnet 5.5（56）仍领先 🔹Argon = Astra = Fable 5.1（53） > GPT-6.1 Sol（52） 🔹折扣价每任务成本只有Astra的60%，图里绿区最香那一格就是它 🔹DeepSWE 77.9%新SOTA，AutomationBench第一，幻觉率同档最低（15%） 🔹输出上限直接开到1M tokens 沉寂7个月后，Google终于把真正的前沿模型端上桌了！先给可信团队，随后API + Ultra。 #Gemini4Argon #GoogleDeepMind #AAIntelligenceIndex #前沿大模型 #Google

This post quotes another post: “Gemini 4 Argon tops Opus 5.5 and GPT-6 Astra on several benchmarks”

x.com/nft_chen/status/2105398642828001777Archived 2026.10.01 10:00