Measured harness ledgerPublic result
DeepSeek V4 FlashMacBook-class cinematic ad scene — DeepSeek V4 Flash Max
Create one polished MacBook-class product-ad shot in Blender with a modeled device, legible industrial detail, intentional materials, lighting, camera movement, and a validator-ready scene.
Max reasoningHeadline result
- Workflow cost
- $0.60
- Wall-clock
- 2h 51m 8.5s wall-clock
- Processed tokens
- 156.27M processed
- Record state
- partial_token_timing_artifact_validation_ledger
Public summary
DeepSeek V4 Flash Max partial_token_timing_artifact_validation_ledger ledger: 2h 51m 8.5s wall-clock, 156.27M processed, and $0.60 API-equivalent usage accounting, not an itemized subscription cash charge.
Run identity and stack
- Result ID: macbook-cinematic-blender-deepseek-v4-flash-max
- Technical model: DeepSeek V4 Flash
- Provider: OpenCode Go
- Stack: OpenCode Go
- Stack: Technical model/configuration: DeepSeek V4 Flash
- Stack: Blender MCP
- Stack: Cinematic product scene
- Stack: Harness v1 MacBook validator
- Stack: Requested tool profile: blender-mcp
Cost basis
- Visible output and reasoning are both output-priced. OpenCode Go subscription pricing is not divided into a marginal cash cost for this run.
Primary artifact integrity
- Kind: blender-cinematic-product-ad-scene
- Path: artifacts/macbook-cinematic-blender-deepseek-v4-flash-max/macbook_cinematic_final.blend
- SHA-256: 6e90f0283efa47577fd02ce6cd1e25750b934cbb3a90465b9df33806e4f1390f
Validation evidence
- Result: PASS, 49/49 checks
- Verification: Supplied result evidence; JSON-validated and matched to the canonical task fixture. It was not freshly rerun headlessly during archival.
- Path: artifacts/macbook-cinematic-blender-deepseek-v4-flash-max/validation_log.json
- SHA-256: 357171055068c541f2012075058f1e8923d20efb60c15c7fef76b5e2459fe9ee
- Validator SHA-256: 54591de8d4ef2953c6f7e2fe3e4f4b3d9407b953049cc858b1012afaf9424d51
Recorded caveats
- This is a two-user-turn protocol deviation, not a one-shot run.
- Wall-clock is end-to-end workflow latency including tool execution, renders, and idle gaps; it is not model-only compute.
- OpenCode splits one agentic turn into multiple assistant/API messages.
- Cache-read tokens are deeply discounted, so total processed tokens overstate cost.
- The public record excludes private OpenCode session/export data; their source hashes and the stable benchmark-window hash are retained.
- The supplied validator log reports PASS 49/49 but was not independently rerun headlessly during archival.
- No render-hardware identity or blind-evaluation record was supplied.
Visible evidence gaps
- one-shot run without an operator rescue turn
- independent headless validator rerun
- render-hardware identity
- 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.