On April 24th, the Zhipu BigModel open platform announced its entry into the "billion-token era," adjusting prices for its various large language models (LLMs) to significantly lower the barrier to entry. This makes large model technology accessible to more businesses at a drastically reduced cost.
This price adjustment affects multiple models on the Zhipu BigModel open platform. The GLM-4-FlashX model now costs only ¥10 per 100 million tokens. Built on a powerful pre-trained base, it boasts ultra-fast inference speeds and functional capabilities comparable to GPT-4, excelling in data extraction, generation, and translation. GLM-Z1-Air and GLM-Z1-AirX are priced at ¥50 and ¥500 per 100 million tokens, respectively. GLM-Z1-AirX's inference speed is over 8 times faster than DeepSeek-R1, making it one of the fastest models in China. GLM-Z1-Air, a cost-effective option, is priced at 1/30th the cost of DeepSeek-R1. Furthermore, the GLM-4-Plus model price has been slashed by 90%, dropping from ¥50 per million tokens to ¥5 per million tokens (equivalent to ¥500 per 100 million tokens), 92% lower than industry average. It offers leading performance in code calculation, data analysis, machine translation, and content generation, meeting the needs of various industries including finance, internet, enterprise services, and education.
Zhipu BigModel's price adjustment aims to empower a wider range of businesses and developers with high-precision models and full-stack technology at a low cost. The platform provides a solid foundation for its entire model suite, ensuring leading model performance without compromising quality. On launch day, over 10,000 businesses and developers accessed the GLM-Z1 series models via bigmodel, with daily token call volume increasing by over 100%, demonstrating strong market demand and recognition.