Unverified50% confidenceFactTime unknown

Haiku 4.5在SWE Bench上得分73.3%，略高于Sonnet 4，接近GPT-5 Codex水平。

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

7/2/2026

First Seen

Valid until: 9/30/2026

Sources

Claude Haiku 4.5 Hands-On: Coding Ability Rivals Sonnet 4 at One-Third the Cost

bilibili三少科技

Related Claims

Partially VerifiedClaude Haiku 4.5在SWE-bench基准测试中取得了73.3%的得分77% similar VerifiedHaiku 4.5计算机使用评分50.7%，超越Sonnet 4的42.2%75% similar Unverified在SWE Bench Pro编程基准测试中，Sonnet 5取得63.2%，Opus 4.8取得69.2%72% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/54470

API

curl https://kongchang.com/api/v1/knowledge/claims/54470

MCP

get_claim(id=54470)