Model Comparison and GLM

💬Community

Thursday, January 8, 2026

Users are comparing Claude Code's performance, particularly with GLM 4.7, against other models like Droid, Minimax, and GPT, while also noting token usage.

3 tweets•0 engagements

@sagacity池建强

Yesterday, I used the integrated Claude Code in Z Code to write a day's worth of code, mainly using GLM-4.7, with the consumed tokens around 6 million. Reading a thousand times is not as good as practicing once. A day's use allows me to have a more intuitive feel for Z Code, such

♥ 80↻ 12

@andyrewleeandyrewlee

tried GLM 4.7 with claude code, opencode and droid. peak experience is with droid.

@archimagosLeonardo Gonzalez

I did a spec-driven cross-model comparison of GLM 4.7, Minimax M2.1, GPT-5.2, and Opus 4.5, holding the agent constant. Once I analyze it I could try choosing a model for a comparison of Droid, opencode, Claude Code, and OpenHands. Or I could try the 4x4 combo free for all