Unverified50% confidenceBenchmarkExact time
同一个DeepSeek v1.1模型在Terminal Bench 2.0上不同Harness下的得分从69.8%跨到74.2%
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/15/2026
First Seen
Valid until: 12/14/2026
Sources
Related Entities
Related Claims
Unverified旗舰Soul在Terminal Bench 2.1基准上取得91.9%的成绩70% similarUnverifiedSWE-2在Terminal-Bench 2.1基准测试中取得了92.8分的成绩67% similarUnverified在SWE-Bench Multilingual测试中，HY3得分75.8，略微落后于DeepSeek V4 Pro的76.264% similarUnverified在DeepSWE上，各模型分数从顶部GPT 5.5的70%到底部Cloud Haiku 4.5的0%，区分度远大于SWE-Bench Pro约30分的集中区间64% similarUnverified根据Artificial Analysis排行榜数据，DeepSeek V4 Pro分数从52分提升到53分，仅有1分差距63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/923846API
curl https://kongchang.com/api/v1/knowledge/claims/923846MCP
get_claim(id=923846)