Unverified50% confidenceBenchmarkExact time

Qwen2.5-Coder和DeepSeek-Coder-V2在HumanEval、MBPP等代码基准测试中的表现已非常接近Claude 3.5 Sonnet等顶级闭源模型

1

Sources

50%

Confidence

Long-term

Relevance

9/1/2026

First Seen

Sources

Cursor接入第三方模型完全指南：摆脱厂商锁定

Related Entities

Qwen2.5-Codermodel DeepSeek-Coder-V2model Claude 3.5 Sonnetmodel

Related Claims

Unverified基准测试显示Qwen3-235B的表现追平甚至超越了DeepSeek R1，在数学、编码、推理等类别中均占优势70% similar UnverifiedDeepSeek-V3发布时在多项基准测试中与Claude 3.5 Sonnet持平甚至超越，训练成本据报道仅为主流闭源模型的数分之一68% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/838972

API

curl https://kongchang.com/api/v1/knowledge/claims/838972

MCP

get_claim(id=838972)