Measured harness ledgerPublic result
Claude Sonnet 5Infinite cathedral corridor shader — Claude Sonnet 5 Max
Generate an infinite stained-glass cathedral corridor shader with convincing depth, architectural repetition, light, and motion.
Max reasoningHeadline result
- Workflow cost
- $18.45
- Wall-clock
- 1h 16m 44.9s wall-clock
- Processed tokens
- 48.10M processed
- Record state
- partial_token_timing_and_shader_ledger
Public summary
Claude Sonnet 5 Max partial_token_timing_and_shader_ledger ledger: 1h 16m 44.9s wall-clock, 48.10M processed, and $18.45 First-party Claude API list-price equivalent, not a subscription invoice.
Run identity and stack
- Result ID: infinite-cathedral-shadertoy-sonnet-5-max
- Technical model: claude-sonnet-5
- Provider: Anthropic Claude Code
- Stack: Anthropic Claude Code
- Stack: Technical model/configuration: claude-sonnet-5
- Stack: ShaderToy fragment shader
- Stack: Harness v1 infinite cathedral prompt
Cost basis
- The aggregate source receipt does not distinguish cache-write TTL; the default 5-minute cache-write rate is applied.
- All-1-hour cache-write alternative: $21.52.
Primary artifact integrity
- Kind: shadertoy-image-pass-fragment-shader
- Path: artifacts/infinite-cathedral-shadertoy-sonnet-5-max/source/infinite-cathedral-shader.glsl
- SHA-256: 6cc1c3171ca53ba84b03b660e89d7babb031f23211a99e591b4651d0c9e18431
Recorded caveats
- Wall-clock is end-to-end latency, not model-only compute; it includes tool-execution time and any gaps between turns.
- The supplied receipt does not report user-turn count, so it cannot be classified as a verified one-shot run from this evidence alone.
- Output tokens include hidden reasoning, generated code, and tool-call JSON, not only visible prose.
- Cache reads dominate processed-token volume but are billed at a deep discount, so processed tokens substantially overstate cost.
- The cache-write TTL is unknown in the aggregate receipt. The displayed $18.4451333 uses 5-minute writes; a 1-hour write assumption would be $21.5188388.
- Static compilation is not a Shadertoy runtime, visual-quality, or frame-rate measurement.
- No initial/optimized local FPS pair, RTX Pro 6000 1080p60 final capture, or blind evaluation is archived.
Visible evidence gaps
- verified one-user-turn benchmark protocol
- initial local FPS
- optimized local FPS
- 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.