Story perspectives
Zhipu AI's GLM-4.7 Surpasses GPT-5 in Coding Abilities
12/23/2025
32 7
1 of 1
Story summary
- Zhipu AI (Chinese start-up) and MiniMax (Chinese start-up) launched flagship open-source AI models within hours to compete for leadership.
- Zhipu AI claims GLM-4.7 has enhanced coding abilities and narrows the gap with Google DeepMind and Anthropic.
- GLM-4.7 surpassed OpenAI's GPT-5 in coding benchmarks and matched Claude Opus 4 on SWE-Bench at over 70%, and Zhipu AI urges users to try it firsthand.
