Expired50% confidenceBenchmarkExact time
The MUSA Coder 27B version outperformed mainstream commercial models such as Claude Opus on the KernelBench evaluation benchmark.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/10/2026
First Seen
Valid until: 10/8/2026(expired)
Sources
GLM-5.2 Coming Soon? Domestic GPU Code LLM MUSA Coder Goes Open Source
bilibili黑鸦Heya6/10/2026
Related Claims
Unverified在SWE-bench Pro基准上，千问3.8 27B模型的分数超过了Claude Opus62% similarUnverifiedQwen3 Coder 30B's performance on programming benchmarks (such as HumanEval and MBPP) is close to or even surpasses some early commercial closed-source models.62% similarUnverifiedFrom the dimension of instruction following, the 27B local model performs closer to the user's original intent than the flagship Opus, but this may simply be because the boundary of its inferential capability is more conservative.59% similarUnverifiedQwen3.6 27B performs close to some commercial models on mainstream benchmarks such as MMLU and HumanEval.57% similarUnverified评测者认为通义千问3.6 27B是30B级别中对高考数学题表现最好的模型56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/469238API
curl https://kongchang.com/api/v1/knowledge/claims/469238MCP
get_claim(id=469238)