Unverified50% confidenceBenchmarkExact time

DeepSeek Coder V1 33B Instruct版本在HumanEval Python代码生成测试中通过率达79.3%，超越GPT-3.5 Turbo

1

Sources

50%

Confidence

Long-term

Relevance

7/10/2026

First Seen

Sources

DeepSeek Coder深度解析：开源代码大模型如何超越GPT-4

bilibiliAI论文许老师教授7/8/2026

Related Claims

VerifiedDeepSeek-Coder的开源版本包括1.3B、6.7B、33B等多种参数规模75% similar VerifiedDeepSeek在多项代码基准测试（如HumanEval、MBPP、LiveCodeBench）中表现优异，部分指标接近甚至超越GPT-4级别模型74% similar UnverifiedDeepSeek-Coder-V2 在 HumanEval 等代码评测榜上长期位居前列73% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/468000

API

curl https://kongchang.com/api/v1/knowledge/claims/468000

MCP

get_claim(id=468000)