Measured harness ledgerPublic result
GLM 5.3 Flash

Neo-gothic storm city shader — GLM 5.3 Flash Max

Generate a TWIGL-style neo-gothic storm city shader with legible atmosphere, architecture, lighting, water, and motion.

Max reasoningHeadline result
Workflow cost
$0.43
Wall-clock
19m 55.549s wall-clock
Processed tokens
467K processed
Record state
partial_multi_turn_static_runtime
Public summary

GLM 5.3 Flash Max partial_multi_turn_static_runtime ledger: 19m 55.549s wall-clock, 467K processed, and $0.43 OpenCode-recorded API-equivalent usage accounting, not itemized subscription marginal cash.

Run identity and stack
  • Result ID: neo-gothic-storm-city-glm-5.3-max
  • Technical model: glm-5.3-flash
  • Provider: OpenCode Go
  • Client: OpenCode
  • Stack: OpenCode Go
  • Stack: OpenCode
  • Stack: Technical model/configuration: glm-5.3-flash
  • Stack: TWIGL fragment shader
  • Stack: Harness v1 neo-gothic storm city prompt
Primary artifact integrity
  • Kind: shader
  • Path: artifacts/neo-gothic-storm-city-glm-5.3-max/source/drowned-gothic.frag
  • SHA-256: f2061556520ce0446baf88899fa2845c5768373b1cb144717ef63407d8effbb7
Validation evidence
  • Result: PARTIAL_TWIGL_UNAVAILABLE
Recorded caveats
  • One exact continue was required before the model produced source, so this is not a one-shot protocol result.
  • TWIGL was not connected; no RTX Pro 6000 render, paired local FPS measurements, or target frame-rate receipt is claimed.
  • The local Chrome capture and static compile are smoke evidence, not blind acceptance evidence.
  • Wall-clock is end-to-end workflow latency, not model-only compute; OpenCode Go does not itemize per-run cash.
  • Private OpenCode receipts, exports, prompts, reasoning, and tool-output text are intentionally not published.
Visible evidence gaps
  • TWIGL-connected execution
  • Initial and optimized local FPS under identical conditions
  • RTX Pro 6000 1080p60 final capture with GPU/frame-rate metadata
  • 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