Z.ai’s open-weights GLM-5.2 beats GPT-5.5 on coding benchmarks

Z.ai’s GLM-5.2 Challenges Frontier Models on Coding Benchmarks at a Fraction of the Cost

Chinese AI startup Z.ai, formerly known as Zhipu AI, released GLM-5.2 as an open-weights model on June 16, delivering a 753-billion-parameter system that outperforms OpenAI’s GPT-5.5 on several long-horizon coding benchmarks while costing roughly one-sixth as much to operate via API.apidog

The model, released under an MIT open-source license on Hugging Face, scored 62.1 on SWE-bench Pro compared to GPT-5.5’s 58.6, and 81.0 on Terminal-Bench 2.1, making it the first open-weights model to cross 80% on that benchmark. On Terminal-Bench, however, it still trails Claude Opus 4.8 at 85.0 and GPT-5.5 at 84.0.venturebeat

Pricing and Access

GLM-5.2’s API pricing stands at $1.40 per million input tokens and $4.40 per million output tokens, with cached input dropping to around $0.26 per million. According to VentureBeat, this works out to approximately one-sixth the cost of GPT-5.5 for comparable coding workloads.llm-stats

The model features a one-million-token context window and supports up to 131,072 output tokens per response. It is already available through Z.ai’s own API, Cloudflare Workers AI, and more than 20 third-party coding environments.ground

Open-Source Competition Intensifies

Z.ai first made GLM-5.2 available to paid coding plan subscribers on June 13 before releasing the full open weights three days later. The model also scored 99.2 on AIME 2026, ahead of both GPT-5.5 and Claude Opus 4.8.marktechpost

The release continues a pattern of increasingly competitive open-source models emerging from Chinese AI labs. GLM-5.2 is the fourth model in the GLM-5 line since February 2026, following GLM-5, GLM-5-Turbo, and GLM-5.1. Z.ai founder Jie Tang framed the release partly as a response to recent restrictions on U.S.-based AI services.gigazine