Unverified60% confidenceOpinionTime unknown
Meta的Llama系列、DeepSeek-Coder以及Qwen是开源阵营缩小与闭源模型代码能力差距的代表
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
6/1/2026
First Seen
Valid until: 8/30/2026
Sources
Qwen在SWE-bench持续领跑：开源AI编程模型的崛起
twitterSWEbench
Related Entities
Related Claims
Unverified开源模型如Llama、Qwen、DeepSeek的崛起，倒逼闭源模型降低API定价以保持竞争力74% similarUnverifiedLlama、Qwen、DeepSeek是当前主要的开源大模型72% similarUnverifiedMeta的Llama系列、DeepSeek-R1、Mistral等主流开源模型大多属于开放权重模型71% similarUnverified作者观察横向覆盖 DeepSeek、GLM、Qwen 等多个开源模型均存在类似工具调用错误模式70% similarVerifiedUnsloth支持微调Gemma 4、Qwen3、DeepSeek、gpt-oss等主流开源大模型70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/26939API
curl https://kongchang.com/api/v1/knowledge/claims/26939MCP
get_claim(id=26939)