AI Model Reviewer

GLM-5.3 in Claude Code 2.1.280

skills off

← Open in the gallerySame prompt, all models

Unfinished output
GLM-5.3's SVG for Isometric SVG town with sprite sheet

This is the unfinished output at the moment the run ended (Harness + model mismatch). See the review below.

Harness + model mismatchIsometric SVG town with sprite sheetmax effortHome harness

Wall time 35m 26sCost $1.16 estimated

Review

Warm autumn countryside town on a thick isometric slab under a hazy sun: about 30 buildings (chapel, town hall with a clock, cottages with smoking chimneys, a greenhouse, a watermill with a wheel, a windmill, a barn with a silo, market stalls) from a 33-symbol sprite sheet, a stepped river with a stone bridge that pours off the edge as a waterfall, dirt roads, a paved plaza with a well and a fountain, lamp posts, sheep, and autumn and pine trees. Z-ordering is clean and the palette is coherent. Faint translucent green discs sit on the grass in a few places.

Verdict by the AI judge (provisional).

The full review

Harness + model mismatch. Reclassified 2026-09-24 (published as ok): the run ended on 'API Error: 400 This model does not support image inputs' when GLM-5.3 read its own render (…/town.png) with Claude Code's Read tool; the artifact was already written but the self-review was cut short (Claude Code x text-only model). Re-run on the text-only relay: isometric_town__claude-code__glm-5.3__attempt2.

Attempt 1 of 3

Every try of this model in this harness on this prompt, oldest first. A later attempt is the cell's result: attempt 3.

  1. Attempt 1 Harness + model mismatch this run35m 26s$1.162026-09-24
  2. Attempt 2 DNF38m 27s$0.842026-09-25
  3. Attempt 3 OK the cell's result46m 16s$3.242026-09-25

Bugs

  • faint translucent green discs on the grass

Scores

Not scored: only runs that finished on their own are scored.

Downloads

Files actually published for this run. Original output, site-made previews and post-run conversions are labelled separately.

The exact prompt

agentic harness (save the file; self-render allowed), 401 characters. Every result for this prompt

Create an isometric SVG town: first design a small sprite sheet of distinct buildings (in <defs>), then place 20+ of them on an isometric grid with roads, trees, a river and a bridge. Pick your own theme and palette. Save it as town.svg in the current directory. You may render and inspect your own output (headless chromium is available at /usr/bin/chromium) and iterate on your own before finishing.

How it was made

Wall time 35m 26sFirst artifact at 34m 16s1 saved version

Versions saved during the run

1 intermediate output, in the order the model wrote them.

Version 1 of 1
  1. Version 1 at 34m 16s thumbnailv1 34:16

Step by step

The prompt, every agent turn and tool call (tool name and a one-line summary; inputs and outputs are not shown), then the final message. The full machine-format transcript is published as a scrubbed download.

Show all 31 steps

Loading the timeline…

Download the full transcript Claude Code session JSONL, 758 KB. Scrubbed: anything that looks like a key, token, e-mail address or phone number is replaced, and account details and the test machine's file paths are removed.

Cost breakdown

Headline cost $1.16, estimated from the token counts at list price.

Tokens by typeTokens$ per 1MCost
Input (uncached)102,088$1.40$0.1429
Cache write0$0.00$0
Cache read1,454,384$0.26$0.3781
Output144,116$4.40$0.6341
of which reasoningnot recordedbilled as output
Total computed from tokens
$1.1552
Total reported by the harness
none

Tokens summed from the Claude Code session JSONL (per message id, incl. subagent files); if Claude Code logged zero usage for this unrecognised model, the /cost screen's per-model usage is used instead. Claude Code tags glm-5p3 as unrecognized_model and its own $ figure has costBasis unknown, so it is NOT used: cost = tokens x Fireworks GLM-5.3 price.

Prices: Fireworks GLM-5.3 serverless price (STATE.md 22:42Z; same as the OpenCode/models.dev catalog entry fireworks-ai/glm-5p3), 2026-09-23.

Run notes

one prompt, agentic harness (self-review allowed)wave 1

  • Mode: interactive TUI in tmux (160x48), inside bubblewrap sandbox
  • Isolation: bwrap: private /tmp, HOME and TMPDIR inside run dir, own CLAUDE_CONFIG_DIR copy; the project files, other runs, /tmp/hx_* not mounted