OpenAI urges four-day workweeks as Z.ai releases top-performing open-source AI model

OpenAI is urging companies to experiment with four-day workweeks as artificial intelligence reshapes the labor market and productivity expectations. According to the BBC, the ChatGPT developer shared these early policy recommendations to spark discussions on adapting to increasingly capable AI systems, emphasizing the need for proactive changes in work structures.^[2]

This push comes amid rapid advancements in AI that could automate routine tasks and boost efficiency, potentially reducing the necessity for traditional five-day schedules. OpenAI's suggestions aim to address how firms can maintain employee well-being while leveraging AI's potential to handle more complex workloads. As reported by the BBC, the company views such trials as a starting point for broader conversations on policy actions required in an AI-driven era.^[2]

The recommendations gain added context from parallel developments in AI safety and innovation. OpenAI recently launched a new fellowship program offering up to $15,000 monthly in AI compute credits, mirroring Anthropic's safety-focused initiative, which underscores the company's commitment to responsible AI scaling, as detailed by Business Insider.^[1] This initiative supports researchers tackling alignment challenges, indirectly tying into workforce adaptation by funding tools that could further transform jobs.

Meanwhile, competition in the AI space intensifies with breakthroughs from Chinese startup Z.ai, known for its GLM family of models. VentureBeat reports that Z.ai released GLM-5.1, an open-source large language model under a permissive MIT License, claiming superior performance over models like Claude Opus 4.6 and GPT-5.4 on benchmarks such as SWE-Bench Pro for software engineering tasks.^[3] Additional sources confirm GLM-5's status as the top open-source performer on leaderboards, scoring 50 on the Artificial Analysis Intelligence Index and excelling in reasoning, coding, and agentic tasks—outpacing rivals like Kimi K2.5 while nearing frontier closed models.^[1][5][6]

These AI leaps highlight why OpenAI's workweek proposal matters: as models like GLM-5.1 enable advanced automation, from long-horizon planning to multi-step workflows with 200K-token contexts, businesses face pressure to rethink human roles.^[2][3] Employees in tech, engineering, and knowledge work stand to be most affected, with potential gains in leisure time but risks of job displacement if transitions lag.

Looking ahead, OpenAI hopes its ideas prompt trials and policy debates, though specifics on implementation remain open. Companies adopting four-day weeks could test productivity metrics against AI-augmented outputs, influencing global labor standards. Z.ai's open-source push, deployable via frameworks like vLLM and SGLang, democratizes access for enterprises to customize and commercialize these tools, accelerating the very changes OpenAI anticipates.^[3][7]

The interplay of policy advocacy and model releases signals a pivotal moment for AI's societal impact, urging stakeholders to balance innovation with equitable work reforms.