Unverified50% confidenceBenchmarkExact time
在Exploit Bench网络安全漏洞分析测试上，SOL表现能与Anthropic的Codex相媲美，但输出Token量仅为后者的三分之一左右
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/11/2026
First Seen
Valid until: 10/9/2026
Sources
GPT-5.6三款模型深度实测：SOL/TERRA/LUNA性能与定价全解析
bilibili薛定猫AI7/9/2026
Related Claims
UnverifiedExploitBench测试中Sol约76%略逊于Fable 5的78%，但Sol仅消耗约12-13万Token而Fable 5用约35万Token63% similarUnverified在Exploit Bench网络安全基准上，Sol得分约76%，Mythos 5约78%62% similarUnverified审查结果显示 Opus 5 频繁杀错进程（常误杀自己所在的 T3 Code 实例），Sol（GPT）有约40%概率提交 draft PR59% similarUnverifiedSol在Terminal Bench 2.1编程测试中获得9019分56% similarUnverified在SWE-Bench基准测试中，Codex得分在69-80分之间55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/488169API
curl https://kongchang.com/api/v1/knowledge/claims/488169MCP
get_claim(id=488169)