🎙️ How I AI: GLM-5.2 review & How Gusto built a new product line with Claude Code

2026-06-29 · Source: Lenny's Newsletter · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Robotics & Autonomous Systems · Depth: Intermediate, extended

Summary

GLM-5.2, an open-weight model from Beijing-based Z.ai, demonstrates production-grade capabilities, benchmarking near Claude Opus 4.8 and above GPT-5.5 on SWE Bench Pro. It features a million-token context window, reasoning mode, function calling, structured output, and context caching. Live testing in a ChatPRD codebase involved audits, UI redesigns, and a 45-minute autonomous bug-hunting task, costing \$3.36 for 6 million tokens. Concurrently, Gusto's CTO, Eddie Kim, detailed how a five-person team built the "Gusto Cofounder" product line in just 10 weeks using Claude Code, bypassing traditional processes like Figma or Jira. This approach highlights how AI, acting as a primary contributor, can accelerate product development from zero code to a tier-one launch, significantly reducing coordination overhead and infrastructure complexity.

Key takeaway

For AI Engineers and ML Directors evaluating model choices, open-weight models like GLM-5.2 present a compelling alternative to expensive frontier models. You should integrate GLM-5.2 into your development rotation for tasks like frontend work or autonomous bug triage, leveraging its competitive performance and significantly lower cost of \$3.36 for 6 million tokens. This strategy reduces vendor lock-in and optimizes operational expenses, but test its consistency with React-heavy workloads.

Key insights

Open-weight models like GLM-5.2 offer production-grade coding capabilities at significantly lower costs, challenging proprietary alternatives.

Principles

• Open-weight models are production-grade alternatives to proprietary LLMs.
• Self-hosting open-weight models reduces vendor dependency and costs.
• AI as a primary team member enables accelerated, process-minimal development.

Method

Configure GLM-5.2 in Cursor by routing an OpenRouter API key, overriding the OpenAI base URL to "openrouter.ai/api/v1/cursor", and adding "z-ai/glm-5.2" as a custom model.

In practice

• Integrate GLM-5.2 for frontend design and long-running agentic backend tasks.
• Design development workflows around AI as a core contributor to minimize overhead.
• Utilize minimal technical stacks (e.g., Cloudflare Workers, Vercel AI SDK) for AI agents.

Topics

• GLM-5.2
• Open-weight Models
• AI Agents
• Claude Code
• Software Development
• Cost Optimization
• Autonomous AI

Best for: NLP Engineer, CTO, VP of Engineering/Data, AI Engineer, Machine Learning Engineer, Director of AI/ML

Related on AIssential

• GLM 5.2: why I’m replacing Opus in Claude Code with this new model — GLM 5.2 offers high-intelligence, cost-effective, open-weight model capabilities comparable to leading proprietary LLMs for coding tasks.
• GLM-5.2 is the step change for open agents — GLM-5.2 is the first open-weight model to credibly rival frontier closed models as a general agent, narrowing the capabilities gap.
• GLM 5.2 is SO GOOD (and almost free) — GLM 5.2 provides high-intelligence, open-weight model capabilities at a fraction of commercial API costs, enabling self-hosting and vendor independence.
• Z.ai’s open-weights GLM-5.2 beats GPT-5.5 on multiple long-horizon coding benchmarks for 1/6th the cost — Z.ai's GLM-5.2 offers a high-performing, cost-effective, and open-source alternative for long-horizon coding tasks.
• TAI #210: GLM-5.2 Closes Most of the Open-Weight Gap in Ten Weeks — GLM-5.2 demonstrates rapid open-weight model advancement through architectural, long-context, and reinforcement learning innovations, closing the gap with frontier models.
• GLM-5.2 First Test | Frontier-Level Open Weight Model? | Agentic Coding, Web Design, Logic Puzzles — GLM 5.2 is a frontier-level open-weight model excelling in agentic coding and complex web design tasks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Lenny's Newsletter.