Hacker News Reader

From the docs

> As advanced models, GLM-5.1, GLM-5 and GLM-5-Turbo are comparable to Claude Opus model. Usage for these models will be calculated and deducted at “3 times the standard rate during peak hours and 2 times the standard rate during off-peak hours”; [...] As a limited-time benefit, GLM-5.1 and GLM-5-Turbo will only consume 1× quota during off-peak hours, valid through the end of April.

https://docs.z.ai/devpack/faq#glm-coding-plan-details

Pelican attempt: https://www.svgviewer.dev/s/8fbPVDUw

Solid but few weird artifacts.

Prompt: > Give me an svg pelican riding a bike

It's probably because the pigeon doesn't know how to ride a bicycle :)) https://www.svgviewer.dev/s/iJuLDemB

Coding evaluation : Claude Opus 4.6 : 47.9 GLM 5.1 : 45.3 GLM 5 : 35.4

What benchmark is "Coding Evaluation"?

Did you work on GLM 5.1?

GLM-5.1 Is Available