Measured harness ledgerPublic result
GLM 5.3 Flash

Space Flight Stage 2 — Landfall — GLM 5.3 Flash Max

Extend the Stage 1 space-flight project with a seamless orbit-to-surface round trip, explorable terrain, takeoff, landing, and a fixed character asset using one frozen follow-on request.

Max reasoningHeadline result
Workflow cost
$19.53
Wall-clock
3h 36m 29.2s wall-clock
Processed tokens
68.12M processed
Record state
partial_multi_turn_token_timing_and_artifact_ledger
Public summary

GLM 5.3 Flash Max partial_multi_turn_token_timing_and_artifact_ledger ledger: 3h 36m 29.2s wall-clock, 68.12M processed, and $19.53 UNVERIFIED reference scenario: GLM-5.2/GLM-5.1 official rates applied to GLM-5.3 token counts; not a verified GLM-5.3 charge or subscription marginal-cash invoice.

Run identity and stack
  • Result ID: space-flight-game-threejs-stage-2-landfall-glm-5.3-max
  • Technical model: glm-5.3-flash
  • Provider: ZCode
  • Stack: ZCode
  • Stack: Technical model/configuration: glm-5.3-flash
  • Stack: Three.js Stage 2 workflow
  • Stack: Harness v1 Landfall prompt
Primary artifact integrity
  • Kind: interactive-threejs-space-flight-stage-2-landfall-game
  • Path: artifacts/space-flight-game-threejs-stage-2-landfall-glm-5.3-max/source/index.html
  • SHA-256: d9a477f0e2b91f1dfde392dd85d662474ee515f2c0da28fc9c9ffa2e4f484c2b
Recorded caveats
  • This is a three-turn, 29-user-message ZCode session, not a one-shot benchmark run.
  • The supplied rollout contains 274 call records while the root-session database reports 275. All token totals reconcile exactly, but the count remains partial.
  • Wall-clock is end-to-end workflow latency, not model-only compute.
  • GLM-5.3 had no official published API pricing at the access date. $19.53324408 is a GLM-5.2/GLM-5.1-rate reference scenario and must not be compared as a verified GLM-5.3 cost.
  • The included images are model-supplied verification shots, not independent RemakeBench replay captures.
  • Static syntax and HTTP serving checks are not a browser/WebGL acceptance, visual-quality, transition-hitch, or FPS measurement.
Visible evidence gaps
  • A clean one-shot generation window
  • Independent browser/WebGL runtime replay of the complete descent, on-foot, boarding, and ascent loop
  • local FPS with browser, hardware, viewport, and transition-hitch receipt
  • RTX Pro 6000 final capture
  • blind-evaluation record
Public result only

This result keeps its public summary and evidence, but it does not currently have a matching Builder test with prompts, projects, Harness workflows or skills.

RemakeBenchResearch console