ModelsThe Decoder· Jun 24, 2026
Snowflake CEO Reports GLM-5.2 Matches Opus 4.7 at Lower Cost
Zhipu AI's GLM-5.2 performed nearly on par with Claude Opus 4.7 in a Snowflake benchmark involving 103 coding tasks. The model achieved this at one-fifth the cost per output token, though it consumed nearly twice as many tokens per task. The pricing difference is described as increasing pressure on Anthropic and OpenAI while potentially affecting valuations of Western AI labs.
Key points
- →GLM-5.2 nearly matches Claude Opus 4.7 on 103 coding tasks
- →Costs one-fifth per output token versus Opus 4.7
- →Consumes nearly twice as many tokens per task
- →Pricing gap pressures Anthropic and OpenAI
Mentioned
SnowflakeZhipu AIGLM-5.2Claude Opus 4.7Anthropic
Related stories
Anthropic and OpenAI Proxy War Over Alex Bores Ends in DrawThe Verge · Policy & Regulation→Pangram CEO Says Language Models Produce Clustered ArgumentsThe Decoder · Models→Claude Tag Integrates Anthropic AI into Slack ChannelsThe Decoder · Product→Mistral AI Releases OCR 4 Model for Document Text ExtractionThe Decoder · Models→Anthropic Launches Claude Tag AI Feature for SlackTechCrunch · Product→ByteDance Unveils Seedance 2.5 Video Model at FORCE ConferenceThe Decoder · Models→
This is an original summary by Dhanasvi's agents based on The Decoder's public feed. For the complete article, visit the original source. Trademarks and article copyright belong to their owners.