GLM 5.3

Zhipu, open weights, released Aug 2026, run at provider defaults through OpenRouter, served by Z.AI. It built the same five pages as every model on season 0.5 of the Rams Index.

Unranked

9 of 10 pages broke: 8 came back with no page and 1 was cut off before it finished. No Landing, Dashboard, Settings or Pricing build finished, and a model needs a scored page on all five briefs to be ranked.

A rank needs every brief, so no model is ranked on the briefs it happened to finish. Its scored pages are shown below, outside the ranking.

Rank
NoneNot every brief has a scored page
Scored pages
1 of 10Broken pages are counted, never scored
Cost per page
$0.14Generation only, every billed build
Cost per scored page
$1.41The same spend over the pages that scored
Season 0.5New this season

How it ran

Every model runs at its provider's default settings, through the same harness, with the same budget per page.

Effort
DefaultThe provider’s default. We send no effort or thinking setting
Time per page
542sMedian over every build, broken ones included
Output tokens
32.0kMedian per page, reasoning included
Reasoning tokens
31.9kMedian per page, as the API reports them
Broken pages
9 of 108 came back with no page, 1 was cut off before it finished. Counted, never scored
Criticals
0%0 of 1 scored pages had a critical issue in their first Rams code review
Built through
OpenRouterServed by Z.AI

Index by brief

The Index on each brief it has a scored page for. A brief where every build broke has none.

Bars run from 40 to 90, the same scale as the ranked models' pages.

The other models

Season 0.5 of the Rams Index, in board order, then the unranked.

Rams Index v0.5 · updated 2026-10-17Judged by Rams eyes · Claude Sonnet 4.6