#glm52

GLM-5.2 self-hosted vs API in 2026: what an open-weight coding agent really costs Summary....

Claude Fable 5 across five agent workloads vs glm-5.2, opus-4-8 and sonnet-5: mid-tool-call refusals, adaptive thinking, and cost that shifts 5-15x by shape.

See my GLM-5.2 open agent benchmark results. It boosted multi-step tool-use reliability by 22% over Mixtral 8x7B in Node.js, slashing hallucinated API calls.

Originally published at norvik.tech Introduction Explore how Z.ai's GLM-5.2 matches...

Z.ai’s GLM‑5.2 Challenges Mythos in Automated Bug‑Finding Zhipu AI has launched the open‑weight...

GLM-5.2 scores 62.1 on SWE-bench Pro vs GPT-5.5's 58.6, ships under MIT, and costs $1.40/M input tokens. Benchmarks, pricing, and the China data question.

Originally published at norvik.tech Introduction Explore how to run the GLM-5.2 model...

On June 13, 2026, the US government cracked down on Anthropic's Claude Fable 5. Hours later, China's ZhipuAI open-sourced GLM 5.2 under MIT license — with a 1M context window and frontier-grade coding scores. This is what happened, why it matters, and how to use it today.

Originally published at norvik.tech Introduction Discover how Z.ai's GLM-5.2...

Set up Z.ai's GLM-5.2 in your IDE: 1M-token span, Anthropic-compatible drop-in, and what the vendor FrontierSWE 74.4 benchmark actually mean

AI Dev Weekly is a Thursday series where I cover the week's most important AI developer news, with my...

I have a confession: I roll my eyes at AI benchmarks. Every other week someone on Twitter posts a...