Unverified85% confidenceFactTime unknown

Claude Haiku 4.5在高中数学竞赛任务中使用工具辅助时得分96.3%，不使用工具时得分80.7%

1

Sources

85%

Confidence

Medium-term (~90 days)

Relevance

6/1/2026

First Seen

Valid until: 8/30/2026

Sources

Claude Haiku 4.5评测：三分之一价格实现旗舰级AI性能

bilibili攒钱换房车的福叔

Related Entities

Claude Haiku 4.5model

Related Claims

UnverifiedClaude Haiku 4.5在研究生水平推理任务中得分73%，多语言问答任务中得分83%，视觉推理任务中得分73.2%76% similar UnverifiedClaude Haiku 4.5在智能体工具使用的零售基准测试中得分83.2%，航空任务得分63.6%，电信任务得分83%75% similar UnverifiedHaiku 4.5的代理工具使用评分达83%，而Claude 4仅为9.6%72% similar Partially VerifiedClaude Haiku 4.5在SWE-bench基准测试中取得了73.3%的得分71% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/24398

API

curl https://kongchang.com/api/v1/knowledge/claims/24398

MCP

get_claim(id=24398)