rbtfl.

US financial/tech media; frames through competition with OpenAI and Anthropic

By lens · 1 takes across the edition

CNBC's June 26 piece, the most-read Western account, reports that GLM-5.2 sits within a percentage point of Anthropic's leading model on agentic benchmarks at roughly one-fifth the cost. Notes Zhipu's stock surged on the announcement and that OpenRouter traffic climbed faster after GLM-5.2 than after DeepSeek V4. The piece frames the release as China's direct answer to the Fable 5 US export ban.

“Zhipu's GLM-5.2 closes to within a percentage point of leading US models on agentic tasks at one-fifth the cost.”