Unverified50% confidenceFactExact time
2023年末开源模型在 SWE-bench 上与 GPT-4 差距显著，闭源模型解决率约为开源模型的2-3倍
1
Sources
50%
Confidence
Long-term
Relevance
7/7/2026
First Seen
Sources
Kimi K2.7 Code Hands-On: How Does High Speed Mode Feel in Practice?
bilibili御风大世界6/23/2026
Related Claims
UnverifiedAt the end of 2023, there was a significant gap in programming capabilities between open-source models and GPT-4; by 2024–2025, domestic open-source models such as DeepSeek Coder, Qwen Coder, and the Kimi K series had substantially narrowed that gap.82% similarUnverifiedSince the second half of 2024, the performance of open-source models has closely approached that of closed-source models, with the ELO scores of Llama, DeepSeek, Zhipu GLM, and other open-source models nearing those of GPT-4o and Claude 3.5.81% similarUnverifiedUsing the release of GPT-4 as a reference point, closed-source models once led the open-source camp by more than a year.78% similarUnverified以Qwen、DeepSeek、Llama为代表的开源/开放权重模型阵营正快速缩小与GPT-4o、Claude等顶级闭源模型的差距73% similarVerifiedGPT-5.5是OpenAI于2025年中期发布的模型，相比GPT-4o在推理深度和上下文理解方面有显著提升72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/168427API
curl https://kongchang.com/api/v1/knowledge/claims/168427MCP
get_claim(id=168427)