Unverified50% confidenceBenchmarkExact time

OpenAI的o系列模型、Anthropic的Claude 3.5 Sonnet在代码生成基准测试（如HumanEval、SWE-bench）上的得分已接近或超过人类程序员平均水平

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

7/10/2026

First Seen

Valid until: 10/8/2026

Sources

FDE是什么？AI落地时代最稀缺的复合型工程师解析

bilibili码士集团-马小鱼7/8/2026

Related Claims

VerifiedClaude Code依托的Sonnet模型在HumanEval、SWE-bench等主流编程基准测试中持续领先77% similar UnverifiedHumanEval由OpenAI自研，专门测试代码生成的功能正确率73% similar Unverified在Pillow库RGB转HSV精度Bug测试中，Claude Code使用Sonnet 4.6模型一次通过，逻辑清晰，获得满分69% similar Unverified测试者使用VS Code Copilot配合OPAI 4.7设计测评方案，用Sonnet 4.6执行测评并生成报告68% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/457477

API

curl https://kongchang.com/api/v1/knowledge/claims/457477

MCP

get_claim(id=457477)