BuilderBuilder
Versioned benchmark collection

Fighter Game v1

Harness v1 · Builder projects v1 · 7 measured runs across 1 target. Each entry is a single recorded attempt; repeat-run reliability is not established.

Inspect the run ledgersOpen Builder reproduction← All collections
Arena-only preview from the clean Fighter Game v1 result
Arena-only preview from the clean Fighter Game v1 result
7Measured runs
1Frozen targets
7Model families
7Configurations
Collection scope

Recorded evidence, without a synthetic score.

Runs share the same frozen target and prompt within each test, with disclosed provider-specific stacks. Costs are recorded estimates as labelled in each ledger, not subscription invoices.

Targets

  • Fighter Game v1

Model coverage

  • DeepSeek V4 Flash · Max
  • DeepSeek V4 Pro · Max
  • GPT-5.6 Luna · Max
  • GPT-5.6 Sol · Max
  • Kimi K3 · Unreported
  • Claude Opus 5 · Max
  • Qwen3.8 Max · xhigh

Evidence available

  • Four selected measured run ledgers
  • Exact frozen prompt identity
  • Per-run arena-geometry artifact identities
  • Exact clean GPT-5.6 Sol rerun with superseded contaminated attempt excluded
  • Prompt SHA-256 672b3d3ed257358e5f7ba780f2f06e2f35194734cad0152c653010a33cd081a6
  • Recorded prompt path benchmarks/fixtures/courtyard-arena-blender-mcp.v1/PROMPT.md

Still missing

  • Blind evaluation and scored side-by-side review
  • Standardized FPS and hitch receipts
  • Repeat-run reliability evidence
Public receipts

7 measured run ledgers

Open any row for its recorded workflow cost, timing, token usage, stack disclosure, artifact integrity, and explicit evidence gaps.

01DeepSeek V4 FlashMax reasoning · Fighter Game v1partial_multi_user_turn_reconciled_metrics_artifact_ledger · Prompt version v1$0.18 OpenCode-recorded, exactly recomputed token subtotal from the supplied rate snapshot; not an actual subscription invoice or marginal cash chargerecorded estimate2h 06m 19.9s wall-clockEnd-to-end workflowOpen 02DeepSeek V4 ProMax reasoning · Fighter Game v1one_shot_metrics_and_artifact_ledger_partial_independent_runtime_evidence · Prompt version v1$0.35 OpenCode-recorded usage accounting, exactly recomputed from the supplied OpenCode Go rate snapshot; not an itemized subscription cash chargerecorded estimate53m 59.574s wall-clockEnd-to-end workflowOpen 03GPT-5.6 LunaMax reasoning · Fighter Game v1partial_multi_user_turn_token_timing_artifact_ledger · Prompt version v1$0.42 API-equivalent estimate from official OpenAI Standard pricing; not the actual subscription-backed Codex chargerecorded estimate33m 29.7s wall-clockEnd-to-end workflowOpen 04GPT-5.6 SolMax reasoning · Fighter Game v1artifact_independent_boot_ledger_closed_book_verified · PASS · Prompt version v1$10.74 API-equivalent estimate from official OpenAI API Standard pricing for the measured token mix; NOT the actual subscription-backed Codex chargerecorded estimate31m 38.6s wall-clockEnd-to-end workflowOpen 05Kimi K3Unreported reasoning · Fighter Game v1artifact_independent_boot_ledger_final_turn_metrics_only · PASS · Prompt version v1≥$2.91 Independently recomputed API-list-price equivalent for the measured token mix; not an itemized subscription charge (lower bound)recorded lower bound31m 15.6s wall-clockEnd-to-end workflowOpen 06Claude Opus 5Max reasoning · Fighter Game v1artifact_independent_boot_metrics_ledger_no_blind_evaluation · PASS · Prompt version v1≥$104.53 First-party Claude API list-price equivalent for the measured token mix; not an itemized subscription charge (lower bound)recorded lower bound1h 23m 18.7s wall-clockEnd-to-end workflowOpen 07Qwen3.8 Maxxhigh reasoning · Fighter Game v1artifact_independent_boot_ledger_partial_metrics · PASS · Prompt version v1¥147.93 Independently recomputed first-party API-list-price equivalent for the measured token mix; NOT a billed amount. This run used an Alibaba Model Studio token-plan subscription, under which usage is not itemized as cash. (upper bound)recorded upper bound1h 17m 10.2s wall-clockEnd-to-end workflowOpen
Prompt provenance

Exact prompt receipts

A path is shown only when the published record includes the prompt hash that identifies it. Campaign evidence and projections keep the hash gap visible.

Separate records

Related records with unknown prompt provenance

These published records stay separate from the prompt receipts above. Unknown provenance is why they are listed here.

No separate unattached records are published for this collection.

Limits

Evidence and material gaps

  • Repeat-run reliability evidence is not established.