Measured harness ledgerPublic result
Claude Fable 5.1Courtyard Blossom Stage 2 — Claude Fable 5.1 Max
Create a second cherry-blossom temple courtyard arena for the fixed fighting game under the pinned reference-guided Stage 2 contract.
Max reasoningHeadline result
- Workflow cost
- $237.58
- Wall-clock
- 2h 35m 15.1s wall-clock
- Processed tokens
- 127.20M processed
- Record state
- artifact_and_sanitized_metrics_with_independent_in_app_browser_replay
Public summary
Claude Fable 5.1 Max artifact_and_sanitized_metrics_with_independent_in_app_browser_replay ledger: 2h 35m 15.1s wall-clock, 127.20M processed, and $237.58 Claude API list-price equivalent, not an itemized Claude Code subscription charge.
Run identity and stack
- Result ID: courtyard-arena-blossom-stage-2-reference-exposed-fable-5.1-max
- Technical model: Claude Fable 5.1
- Provider: Anthropic Claude Code
- Stack: Anthropic Claude Code
- Stack: Technical model/configuration: Claude Fable 5.1
- Stack: Blender / browser arena workflow
- Stack: Reference-guided Stage 2 fixture
Primary artifact integrity
- Kind: manifest-verified-artifact
- Path: artifacts/courtyard-arena-blossom-stage-2-reference-exposed-fable-5.1-max/source/scripts/arena_blossom/arena_blossom.blend
- SHA-256: f1a8e6a056e3582510ae56ace6df216bd148740cbf7c6d1569d6b45bdc91900c
Recorded caveats
- Wall-clock is end-to-end workflow latency, including tool execution and user idle time, not model-only compute.
- Output tokens include hidden reasoning, generated code, and tool-call JSON in addition to visible text.
- Cache reads are discounted to 2.5% of fresh-input rate for the supplied Fable 5.1 pricing record, so processed-token volume substantially overstates effective cost.
- The original exact post-generation script selected another model's newer subagent log by mtime; this public record uses the corrected Fable workflow ledger and binds its raw receipt by hash while withholding private identifiers.
- The retained source contains only privacy-normalized copies of the builder and Blender inspection renders: private absolute path metadata was removed, the builder has a portable project-root fallback, and 46 inspection renders had zero absolute pixel error against their supplied counterparts.
- The archive operator's replay proves the two routes and real start input, but does not establish cross-hardware FPS, transition hitch, or visual quality.
- No blind evaluation or standardized performance receipt is recorded.
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.