Unverified50% confidenceBenchmarkExact time
2024-2025年间，Claude 3.5 Sonnet、GPT-4o、Gemini 2.0等模型在HumanEval、SWE-bench等编程基准测试的通过率从不足50%提升至90%以上
1
Sources
50%
Confidence
Long-term
Relevance
8/23/2026
First Seen
Sources
Vibe Coding进阶：从玩具项目到企业级AI编程实战指南
Related Entities
Related Claims
VerifiedIn 2024, models such as Claude 3.5 Sonnet have already exceeded 50% on the SWE-bench Verified subset.75% similarUnverifiedClaude 3.5 Sonnet在HumanEval基准测试上的得分超过90%，而2021年最初的Codex模型得分仅为28.8%74% similarUnverifiedIn SWE-bench testing, the early GPT-4 resolution rate was below 2%, while Claude 3.5 Sonnet broke through 49%.73% similarUnverifiedClaude Sonnet 4.6 improved from 61.4% to 72.5% on the OS World (agentic computer use) test.70% similarUnverifiedOn mainstream benchmarks such as MMLU and HumanEval, the score gaps among Claude 3.5, GPT-4o, Gemini 1.5 Pro, and Grok-2 have been compressed to within the range of statistical error.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/789977API
curl https://kongchang.com/api/v1/knowledge/claims/789977MCP
get_claim(id=789977)