GLM-5.1 from Z-AI is live on Chat-O. It’s a focused upgrade over GLM-5 in one specific area: handling long-horizon tasks — multi-step coding jobs where a model has to stay coherent across many rounds of work rather than producing a single one-shot answer.

What GLM-5.1 improves over GLM-5

• Long-horizon coding — fewer “lost the plot” failures on multi-step tasks
• Large context handling — better use of the full 200K-token window
• Reasoning stability — more predictable chain-of-thought across long prompts

Z-AI specifically calls out long-horizon tasks as a chief area of improvement, which made GLM-5.1 a natural fit for the Power tier on Chat-O.

Where it fits

GLM-5.1 is the choice when your task is too long for GLM-5 to stay focused but doesn’t need the full 1M-token window of GLM-5.2.

Try it

Pick GLM-5.1 in the model dropdown on any chat. Power-tier credits apply — if you’re on the free tier, you still get trial credits to test it.