Unverified50% confidenceBenchmarkExact time
Argon在CWE Bench上以68%与GPT-6 Astra、Grok 4.7并列第一
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/3/2026
First Seen
Valid until: 1/1/2027
Sources
Related Entities
Related Claims
Unverified在 Deep Study 软件工程测试中,Argon 得分 77.9%,高于 GPT-6 Astra 和 Cloud Opus 5.5(后两者约为 74%)75% similarUnverified在 Frontier Software Engineering 基准上 GPT-6 Astra 以 65.5% 领先,Opus 5.5 为 62.3%,Gemini 4 Argon 仅 55%75% similarUnverifiedGrok 4.5在Terminal Bench基准测试中取得83.3%,基本与GPT 5.5打平,略低于Claude 567% similarVerifiedGPT-6 Astra的API模型标识为gpt-6-astra66% similarUnverifiedGPT-6 Astra在KingBench 3基准上获得90%的高分(72/80)64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/969182API
curl https://kongchang.com/api/v1/knowledge/claims/969182MCP
get_claim(id=969182)