gemini-4POST
Gemini 4 Argon rivals Opus 5.5 in plain-language output, one user says
A Chinese user says Argon's plain-language ability is close to Opus 5.5 and its instruction following beats GPT-6.1 Sol and Astra, with looser guardrails, and may become their main model.
Sn1waR
@SnowyWar36965Archived 20 days agoOriginally posted 2026.10.14
Sn1waR @SnowyWar36965Originally posted 2026.10.14
Gemini 4 Argon高强度体验了一下，说人话能力跟opus5.5差不多， 指令理解和执行这块比gpt6.1sol和astra好太多了，没有那么多傻逼边界护栏钟墙门禁。跟opus5.5差距暂时没体验出来，但是胜在没有cyber墙（拿来改muteki没触发过云护栏拦截。） 可能接下来是我的主力模型。gpt越来越难用了。。。
x.com/snowywar36965/status/2110055323361067030Archived 2026.10.14 10:02