Unverified50% confidenceBenchmarkExact time
Qwen2.5-Coder 系列由阿里巴巴通义实验室出品，其32B参数版本在多项指标上超越了GPT-4o
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/21/2026
First Seen
Valid until: 10/19/2026
Sources
Related Claims
Unverified阿里云Qwen2.5-Coder-32B在2024年底的评测中，在HumanEval、MBPP等主流代码基准上超越GPT-4o82% similarUnverifiedQwen3.5在HLE博士级综合题评测中得分28.7，而GPT-5.2为35.573% similarUnverifiedQwen3-4B在多项推理类评测指标上已与DeepSeek V3或GPT-4o相差无几72% similarVerifiedQwen3.6系列模型已经发布，包含27B和35B MOE等不同规格72% similarUnverified阿里开源了Qwen3.5，这是一个397B总参数、仅激活17B参数的混合专家模型72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/579916API
curl https://kongchang.com/api/v1/knowledge/claims/579916MCP
get_claim(id=579916)