Measured harness ledgerPublic result
Gemini 3.6 FlashInfinite cathedral corridor shader — Gemini 3.6 Flash High · Attempt 1
Generate an infinite stained-glass cathedral corridor shader with convincing depth, architectural repetition, light, and motion.
High reasoningHeadline result
- Workflow cost
- $1.92
- Wall-clock
- 7m 09s wall-clock
- Processed tokens
- 3.71M processed
- Record state
- published_artifact_runtime_ledger
Public summary
Gemini 3.6 Flash High published_artifact_runtime_ledger ledger: 7m 09s wall-clock, 3.71M processed, and $1.92 Google Cloud Agent Platform Standard Gemini 3.6 Flash API-equivalent scenario; not an Antigravity Starter invoice.
Run identity and stack
- Result ID: infinite-cathedral-shadertoy-gemini-3.6-flash-high-antigravity-cli-attempt-1
- Technical model: gemini-3.6-flash
- Provider: Google Gemini via Antigravity
- Client: Antigravity CLI 1.1.5
- Attempt: 1
- Stack: Google Gemini via Antigravity
- Stack: Antigravity CLI 1.1.5
- Stack: Technical model/configuration: gemini-3.6-flash
- Stack: Attempt 1
- Stack: ShaderToy fragment shader
- Stack: Harness v1 infinite cathedral prompt
Cost basis
- actual incremental antigravity cash charge usd: $0.00 USD.
Primary artifact integrity
- Kind: shadertoy-fragment-shader
- Path: artifacts/infinite-cathedral-shadertoy-gemini-3.6-flash-high-antigravity-cli-attempt-1/source/shadertoy_code.glsl
- SHA-256: 162aaa7c7a75e2aae04ed5df47a5b387ee457c9f86b256577188ce9c11c9c1ed
Recorded caveats
- Wall-clock is end-to-end latency, including approvals and tool execution, not model-only compute time.
- Output already includes the 27,920 reasoning tokens; the visible-response and reasoning values are a decomposition, not additive categories.
- The public dollar amount is an API-equivalent comparison, not a subscription charge.
- The raw transcript, client log, quota screenshot, and per-request usage receipt are excluded from the public repository because they contain model thinking, user identity, conversation/database identifiers, or local paths; source hashes and sanitized aggregate receipts are retained.
- The 75-FPS handoff observation conflicts with the retained frame's self-reported 3-FPS badge, and neither is comparable to the benchmark's standardized shader performance workflow.
- No formal RemakeBench quality score or blind-voting result has been recorded.
Visible evidence gaps
- initial local FPS measurement receipt
- optimized local FPS measurement receipt
- RTX Pro 6000 1080p60 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.
