Unverified60% confidenceBenchmarkExact time
GPT-5.6在Terminal-Bench 2.1编码榜单上以Ultra配置取得91.9%的成绩
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
7/12/2026
First Seen
Valid until: 10/10/2026
Sources
GPT-5.6发布：编码跑分反超竞品，推理速度提升10倍
bilibili天选嘉鲤敦赵图图7/10/2026
Related Claims
UnverifiedGPT-5.6 Sol标准模式在Terminal Bench 2.1上超越Claude Mythos 5，开启Ultra Mode后成绩达到91.9%80% similarUnverifiedGPT-5.6编程能力在关键基准上达到91.9%，安全评测CyberGym得分85.6%，均超越竞品Nexus 579% similarUnverifiedGPT 5.5在TerminalBench 2.0基准测试中以87%的成绩大幅领先，超过了Anthropic尚未发布的Mythos模型78% similarPartially Verified三个 GPT-5.6 模型在长程代理微调任务上均拿满分 10 分，而 GPT-5.5 和 Opus 4.7 在该任务只能拿两三分78% similarUnverifiedGPT-5.6 Sol在BrowseComp浏览任务基准上达到92.2%78% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/493632API
curl https://kongchang.com/api/v1/knowledge/claims/493632MCP
get_claim(id=493632)