GLM Ships 5.1 Open Source LLM, Outperforms Opus 4.6 and GPT-5.4 on SWE-Bench Pro

Summary

GLM has released its 5.1 open source large language model, which reportedly outperforms both Anthropic's Opus 4.6 and OpenAI's GPT-5.4 on SWE-Bench Pro, a benchmark focused on real-world software engineering tasks. The headline framing of AI "joining the 8-hour work day" suggests the model demonstrates sustained, reliable performance on extended coding and engineering workflows comparable to a full working session. This is a significant development for the open source model ecosystem, as an ope